Servo2008 03

V
o
l
.

6

N
o
.

3
S
E
R
V
O
M
A
G
A
Z
I
N
E
R
O
B
O
T

K
I
T
S
G
E
N
E
T
I
C

A
L
G
O
R
I
T
H
M
S
W
O
M
E
N

O
F

R
O
B
O
T
I
C
S
B
I
O
L
O
G
I
C
A
L

C
O
N
T
R
O
L

S
I
G
N
A
L
S
M
a
r
c
h

2
0
0
8
Cover.qxd 2/6/2008 2:16 PM Page 1
Full Page.qxd 2/4/2008 10:07 AM Page 2
This Month In
THE COMBAT ZONE ...
Features
28 The Holy Grail of Combat Robotics
Usable Melty Brain (Part 2)
31 Manufacturing: Win With
Bulletproof Planetary Gearboxes
Events
34 Results and Upcoming Competitions
35 Rumble at the Rock:
BotsIQ Gone Varsity
Robot Profile
36 Dark Pounder
SERVO Magazine (ISSN 1546-0592/CDN Pub Agree#40702530)
is published monthly for $24.95 per year by T & L Publications, Inc., 430
Princeland Court, Corona, CA 92879. PERIODICALS POSTAGE PAID
AT CORONA, CA AND AT ADDITIONAL ENTRY MAILING OFFICES.
POSTMASTER: Send address changes to SERVO Magazine, P.O.
Box 15277, North Hollywood, CA 91615 or Station A, P.O.
Box 54, Windsor ON N9A 6J5; cpcreturns@servomagazine.com
Departments
06 Mind/Iron
19 Events Calendar
26 New Products
27 Robotics Showcase
65 Menagerie
66 Robo-Links
75 SERVO Webstore
82 Advertisers Index
Columns
08
Robytes by Jeff Eckert
Stimulating Robot Tidbits
10
GeerHead by David Geer
Pace Robotics Lab Activevision
Robot Technology Captures
Sights in 3D
14
Twin Tweaks
by Bryce and Evan Woolley
Back to Basics: Why Turning RC Cars
into Robots Makes All the Difference
20
Ask Mr. Roboto by Pete Miles
Your Problems Solved Here
67
Dif ferent Bits
by Heather Dewey-Hagborg
Artificial Life Part 1:
Introduction to Genetic Algorithms
72
Robotics Resources
by Gordon McComb
Robot Kits for Easier Robotics
78
Appetizer
by Allison F. Walton and Filomena Serpa
When Art and Servos Mix
79
Then and Now by Tom Carroll
Women of Robotics
PAGE 79
4 SERVO 03.2008
TOC Mar08.qxd 2/6/2008 10:32 AM Page 4
03.2008
VOL. 6 NO. 3
SERVO 03.2008 5
38 Control of Power-Assist
Exoskeleton Robots With
Biological Signals
by Kzauo Kiguchi
Take a look at one soft computing
technology that hopes to help bring
exoskeleton robots to consumers.
43 Reviving an Androbot BOB
by Robert Doerr
This time, BOB gets lots of cool stuf f
added to him, including a Handy
Board controller, H-bridge, power
distribution, and sonar boards.
50 Building a Stepper Motor
Controller: Part 2
by Fred Eady
The preflight work from last month
will now be applied to create
controlled rotational movement
of a stepper motor shaft.
58 A More Versatile Robotic
Controller Using
RobotBASIC
by John Blankenship and
Samuel Mishal
RobotBASIC is a free programming
language known for its integrated
robot simulator. However, it also
can be used for nearly any control
application.
62 Designing and Building a
Robot From Scratch
by Brian Benson
This new series of articles will
take you through the entire
process of designing and
building a custom robot.
PAGE 38
PAGE
10
PAGE 14
Features & Projects
TOC Mar08.qxd 2/6/2008 10:33 AM Page 5
Published Monthly By
T & L Publications, Inc.
430 Princeland Court
Corona, CA 92879-1300
(951) 371-8497
FAX (951) 371-3052
Webstore Only 1-800-783-4624
www.servomagazine.com
Subscriptions
Toll Free 1-877-525-2539
Outside US 1-818-487-4545
P.O. Box 15277
North Hollywood, CA 91615
PUBLISHER
Larry Lemieux
publisher@servomagazine.com
ASSOCIATE PUBLISHER/
VP OF SALES/MARKETING
Robin Lemieux
display@servomagazine.com
EDITOR
Bryan Bergeron
techedit-servo@yahoo.com
CONTRIBUTING EDITORS
Jeff Eckert Tom Carroll
Gordon McComb David Geer
Pete Miles R. Steven Rainwater
Fred Eady Kevin Berry
John Blankenship Samuel Mishal
Kazuo Kiguchi Robert Doerr
Brian Benson Allison Walton
Nick Martin Michael Bastoni
Bryce Woolley Evan Woolley
Heather Dewey-Hagborg Filomena Serpa
CIRCULATION DIRECTOR
Tracy Kerley
subscribe@servomagazine.com
MARKETING COORDINATOR
WEBSTORE
Brian Kirkpatrick
sales@servomagazine.com
WEB CONTENT
Michael Kaudze
website@servomagazine.com
PRODUCTION/GRAPHICS
Shannon Lemieux
Joe Keungmanivong
ADMINISTRATIVE ASSISTANT
Debbie Stauffacher
Copyright 2008 by
T & L Publications, Inc.
All Rights Reserved
All advertising is subject to publishers approval.
We are not responsible for mistakes, misprints,
or typographical errors. SERVO Magazine
assumes no responsibility for the availability or
condition of advertised items or for the honesty
of the advertiser. The publisher makes no claims
for the legality of any item advertised in SERVO.
This is the sole responsibility of the advertiser.
Advertisers and their agencies agree to
indemnify and protect the publisher from any
and all claims, action, or expense arising from
advertising placed in SERVO. Please send all
editorial correspondence, UPS, overnight mail,
and artwork to: 430 Princeland Court,
Corona, CA 92879.
As highlighted by the Consumer
Electronics Show (CES), 2008 is
shaping up to be a good year for
robotics innovators who rely on
inexpensive, capable platforms for
their experiments. The most impressive
platform in the traditional D3 (dull,
dirty, and dangerous) camp was the
$99 iRobot Looj (www.iRobot.com).
The treaded, weather resistant, remote
controlled vehicle (see Figure 1) is
intended to facilitate the D3 job of
cleaning gutters of leaves, pine cones,
twigs, and other light debris. Although
the robots movement is limited to
linear forward and reverse and access
to the NiCad battery pack is somewhat
awkward, the robot seems like the
perfect platform for an amphibious
vehicle. Ill be featuring the Looj in an
upcoming teardown article.
As in previous years, consumer
robots is less about sweeping floors,
clearing gutters, or performing equally
distasteful D3 tasks, and more about
entertainment. WowWee (www.wow
wee.com) promises to be a ready
source of development platforms,
following the popularity of the
Robosapien among the robotics mod
community. Highest on my want list is
the Bladestar indoor flying machine
with onboard obstacle avoidance. The
three-channel digital IR controller is
especially intriguing, in that it implies
control is possible through a wireless
laptop link that can be used to provide
path planning in addition to simple
object avoidance. Other members of
the WowWee lineup including the
four-legged Roboquad, emotive
Robopanda, and series of Alive pets
also seem promising.
If youve ever developed a WiFi
interface to your robots, then youll
appreciate the Meccano Spykee spy
Mind / Iron
by Bryan Bergeron, Editor
Mind/Iron Continued
6 SERVO 03.2008
FIGURE 1
Mind-Iron Mar08.qxd 2/6/2008 10:55 AM Page 6
robot with built-in Skype VoIP phone, webcam, and
software suite (www.meccano.com). The French robot,
sold under the Erector brand in the US, seems equally
valuable as a source of parts and as a development platform.
At $300, the treaded robot is about the price of a WiFi
webcam without audio capabilities, battery pack, or mobility.
Im undecided about the viability of the long-awaited
Pleo (www.pleoworld.com) as a repurposable robotics
platform. Given the hype, I was expecting something with
the capabilities of the discontinued Sony Aibo. However, I
dont envision squads of autonomous, bucolic Pleos
playing robosoccer. That may change with the efforts of an
innovative modder, however.
One of the more interesting robotics products
featured at the 2008 CES that spans the D3 and
entertainment categories is the Gibson Robot Guitar
(www.Gibson.com). Thanks to robotics and electronics
developed by Tronical (www.Tronical.com), the guitar
frees the guitarist from the dull and time-consuming task
of retuning the guitar. Its difficult to rationalize the added
$700 expense for simply keeping six strings in tune, but
where the German Tronical technology shines is in
alternative tuning and intonation adjustments.
Many traditional and modern songs use alternatives to
the standard EBGDAE tuning (i.e., the first or thinnest
string is tuned to E, the next string to B, and so on).
Retuning a guitar to common alternatives such as Dropped
D (DADGBE) or Delta Blues (DGDGBD) takes time. So much
time that performers typically switch guitars between
songs to accommodate alternative tuning. With the Robot
Guitar, alternative tuning is as simple as turning a selector
switch. The six motors in the head and piezo audio
detectors in the bridge adjust individual strings to the
appropriate tension within four seconds.
The other big headache the Tronical technology
addresses is adjusting intonation, which typically involves a
trip to the guitar shop for adjusting the bridge. Instead of
simply pressing a button, correcting intonation involves
manually adjusting the distance between points holding
the strings (the nut at the far end and the bridge at the
near end). While this is a manual operation, the circuitry in
the guitar signals the operator how far to turn the
adjustment screws to achieve proper intonation.
There have been other automatic, motorized tuners on
the market, but the Tronical-Gibson is the first to pull it off
in a clean, fully integrated way. Because the Tronical
components are the same size and actually lighter than
traditional components, the system is available for Fender
Strats and a variety of Gibson guitars. The system can be
quickly installed and removed tracelessly without extra
drilling, holes, or screws.
While the market for all stringed instruments is
threatened by all electronic instruments, the Robot Guitar
is a good example of how robotics can be integrated
seamlessly and almost invisibly into an existing product to
provide enhanced value. I leave you with the challenge of
identifying application areas where the same approach can
be applied to activities of daily living, from driving and
cooking to simply moving from one place to another. SV
SERVO 03.2008 7
Extreme Robot Speed Control!
OSMC
6
6
6
6
Monster power!
14-50V
3.15 x4.5x1.5
3 wire interface
160A!
www.robotpower.com
Phone: 253-843-2504 sales@robotpower.com
Scorpion HX
6
6
6
6
Dual (6A pk) H-bridges
fwd-only channel
5V - 18V
1.6 x 1.6 x 0.5
2.5A
12A Plus
$79.99
Scorpion Mini
6
6
6
2.5A (6A pk) H-bridge
5V - 18V
1.25 x 0.5 x 0.25
$119.99
Scorpion XL
6
6
6
Dual H-bridge
5V - 24V
2.7 x 1.6 x 0.5
13A 45A Peak!
6
6
6
6
6
6
6
14V - 50V - Dual H-bridges - 150
Adjustable current limiting
Temperature limiting
Three R/C inputs - serial option
Many mixing options - Flipped Bot Input
Rugged extruded Aluminum case
4.25" x 3.23" x 1.1
80A A+ Peak!
$29.99
$399
Introducing Dalf
6
6
6
6
6
6
6
Closed-loop control of two motors
Full PID position/velocity loop
Trapezoidal path generator
PIC18F6722 CPU
C source for routines provided
See
Giant Servo Mode!
www.embeddedelectronics.net
$250
MADE IN
THE USA
H-bridges: Use with Dalf or with your Micro/Stamp
Simple-H
6
6
6
6
6-28V 25
2.25 x2.5 x0.5
3 wire interface
current & temp
protection
A!

$79
Mind-Iron Mar08.qxd 2/6/2008 10:56 AM Page 7
8 SERVO 03.2008
Climbing the Walls
In the common tradition of borrow-
ing robotic concepts from nature is
Waalbot, which needs no magnets or
vacuum devices to attach itself to
vertical planes. Like a common gecko,
this Carnegie Mellon (www.cmu.edu)
invention uses tiny fibers on its feet to
adhere to just about any surface. The lit-
tle guy isnt much bigger than a quarter,
but he sports two sets of three-footed
wheels, each with its own motor. The
spring-loaded tail keeps the critter
pushing against the walls surface.
Motion control, including steer-
ing, is provided by a PIC microcon-
troller and power by lithium-ion
batteries. Projected applications
include inspection, surveillance, and
possibly spacecraft repair. Coming
soon to a Waal-mart near you.
Dinosaurs to Roam Again
Dubai, to put things in perspec-
tive, is the second largest nation within
the United Arab Emirates, even though
it occupies only 4,114 sq km (about 16
sq mi). This puts it on a par with
Headland, AL. But its population is
1,422,000, as opposed to Headlands
3,523, so a lot of people must be
standing up most of the time. On the
positive side, Dubais gross domestic
product in 2006 was $46 billion, which
means they have a lot of extra money
for fun projects. And these folks, who
have already created a private island
archipelago shaped like the Earths
continents, and the worlds first
underwater hotel, dont think small.
The current hot project there is
Restless Planet, a unique, world-class
natural history phenomenon that will
recreate 11 acres of the Earth as it was
100 million years ago. The park
projected to cost $1.1 billion will
feature 109 robots housed in a 75 m
dome, constituting the worlds largest
collection of animatronic dinosaurs.
The bots are being created by Japans
Kokoro Co. under the direction of
famed paleontologist Jack Horner.
The first one out of the gate is T.
Rex (the lizard, not Marc Bolan), which
is capable of follow-
ing you with hungry
eyes, breathing, and
curling its lips, but it
will probably stop
short of eating you.
A series of rides will
take visitors through
a collection of
high-tech effects that
illustrate the birth of
the planet and the
creation of its topographical features
and oceans. The finale is a visit to the
age of dinosaurs. Restless Planet is
scheduled to open late this year, so
book your flight to the City of Arabia
(www.cityofarabiame.com) early. (The
current price is $1152, round trip KLM.)
Baby Seals Boost
Mental/Physical Health
Most of the bots you see these days
are aimed at some sort of mundane
application, be it industrial or service. But
Japanese developers seem to be wrapped
up in what has been called the cult of
cute, and one of the most adorable is
Paro, the baby harp seal from Intelligent
System Co. (www.intelligent-system.
jp). It is classified as a mental commit-
ment robot, defined as one devel-
The CMU Waalbot climbs walls using
dry adhesion. Photo courtesy of
Carnegie Mellon Nanorobotics Lab.
Left: Paro, the robotic baby harp seal, photo courtesy
of Intelligent System Co., Ltd. Right:The real thing,
photo by Rei Ohara, courtesy of harpseals.org.
The Restless Planet theme park will feature 100+ mechatronic
dinosaurs. Photo courtesy of City of Arabia.
by Jeff Eckert
Robytes.qxd 2/5/2008 6:02 PM Page 8
oped to interact with human beings
(often the sick and elderly) and make
them feel emotionally attached to it.
According to the company, such
devices provide three basic therapeutic
effects: psychological (e.g., relaxation
and motivation), physiological
(improved vital signs), and social
(stimulating communication between
patients and caregivers). Going beyond
a common stuffed animal, Paro
incorporates tactile, light, audio, tem-
perature, and posture sensors to com-
prehend people and its environment.
It recognizes light and dark and
gets sleepy at night. It blinks its eyes
and makes seal noises. It likes to be pet-
ted and tries to avoid you if you smack
it. Paro can even recognize words and
where your voice is coming from, and
you can tickle it by touching its whiskers.
Pretty clever. But with Paros
$3,200 price tag, a cat might be a
more cost-effective solution.
Dont Eat the Yellow Ice
Also more huggable than it needs
to be is Yuki-taro, from Research &
Development, Inc. (RDI, www.rdi-
japan.com), which has been described
as a supercute robot that eats up
snow and poops ice blocks. Developed
by a consortium in Japans snowy
Niigata Prefecture, it is self-guided via a
GPS system and cameras mounted in
his eyes. He measures 63 x 37 x 30 in
(160 x 95 x 75 cm), weighs in at 880 lb
(400 kg), and his droppings are 24 x 12
x 6 in (~60 x 30 x 15 cm) ice bricks.
Given the nature of his diet, you
probably wont want to crush up the
bricks for your evening cocktail, but
the ice could be stored for refrigera-
tion or air conditioning in summer
months. Yuki-taro isnt ready for mass
production yet, but its inventors hope
to be selling them within five years.
The estimated price will be $9,000.
Report on Future Military
Systems
In its infinite wisdom, the US
Department of Defense (www.
defenselink.mil) has released a report
titled, Unmanned Systems Roadmap
20072032, which outlines how
the military intends to proceed in
developing, acquiring, and integrating
unmanned technology over the next
25 years. This should prove helpful to
inventors, defense contractors, hostile
governments, terrorists, and anyone
else who has an interest in such things.
The Roadmap covers not only
UAVs but land- and maritime-operated
systems, as well. The report is available
at a somewhat out-of-the-way corner
of the DoD website, or you can
download it at www.jkeckert.com/
unmannedsystems.pdf.
Do-It-Yourself Earth Defense
Strangely, the DoDs Roadmap
completely ignores the threat of aliens
from outer space, but Daniel H. Wilson,
Ph.D., has it covered in his latest book,
How to Build a Robot Army. Regardless
of whether you find yourself attacked
by aliens, ninjas, or zombies (or mum-
mies or great white sharks or Godzilla),
you can get the better of your blood-
thirsty adversaries using the techniques
outlined within. You can pick up the
paperback edition from amazon.com
for a paltry $11.16 as of this writing.
And may the farce be with you. SV
Robyt es
Yuki-taro, the snow-eating robot.
Photo courtesy of RDI.
New report from the DoD outlines its
plans for future unmanned systems.
Photo courtesy of
US Department of Defense.
Daniel Wilsons latest book,
How to Build a Robot Army.
Photo courtesy of Bloomsbury USA.
SERVO 03.2008 9
Robytes.qxd 2/5/2008 6:02 PM Page 9
10 SERVO 03.2008
P
ace University Labs produced the
activevision technology (per a
Pace University academic paper)
in conjunction with research into a
much larger robot cognition project.
With activevision, the robot models
itself and its environment in a 3D world
using graphics rendering engine
technology from Ogre3D, just like that
used in gaming software.
The robot sees the world around it,
then assembles it in 3D. It saves and
works within that reservoir of graphical
data in order to develop changing and
improving perceptions of its surroundings.
ADAPT-ing
The robot vision project, called
Adaptive Dynamics and Active
Perception for Thought (ADAPT), falls
under the work of three University
research groups: one from Pace
University (computer science), one from
Brigham Young, and one from Fordham.
The research has produced several
robots, which are available from
ActivMedia today along with some
others. These robots are capable of
a variety of responses in largely
unpredictable environments using
robotic cognition and activevision.
There were obstacles to accomplish-
ing this level of perception with earlier
robots. Developers had to pre-program
those robots to work in their environ-
ments. So, while the pre-programming
had a lot to do with how they could
respond, it didnt help them learn from
the environment or produce their own
perceptions before they responded.
By developing robotic cognitive abil-
ities, researchers hope to be able to give
robots the tools they need to learn from
their environments and adapt according-
ly. Some of the pieces to that puzzle
include the abilities for the robot to solve
problems and improve navigation.
Seeing is Predicting
The mobile robot in the ADAPT
research sees by first predicting what it
will see. It does this using a virtual
model of the world around it through
its multimedia. This exists in the SOAR
software and algorithms (SOAR is a
cognitive software architecture and
framework for developing intelligent
robotics systems), and in memory.
The virtual, multimedia aspect of
the model exists in a 3D game.
Ogre3D is an open-source programming
platform, virtual-world-based game with
advanced graphics. It uses state-of-the-
art game physics (the physics make virtu-
al objects in the game respond to each
other in the same way that the same
objects would respond in the real world).
The robot uses its machine vision
and software tools to create a copy of its
environment with itself included. It stores
and interacts with 3D data in the virtual
game world, learning from the process.
This forces the robot to use its intelli-
gence so it can make decisions based on
its perceptions of the world around it and
not on feedback (based on machine
vision techniques) from that world alone.
The robot learns to adapt and
respond to the world around it as a part
of a complex problem-solving process.
The robots software uses the virtual
Contact the author at geercom@alltel.net by David Geer
PACE ROBOTICS LAB
Activevision Robot Technology Captures Sights in 3D
Pace Robotics Lab has developed a real looker a robot with
machine vision that remembers the world in 3D.
ActivMedia Activrobot, Pioneer 2 rear side-
angle view. This early experimental version is
the granddaddy to ActivRobots Pioneer 3
models, including the P3-DX. The P3-DX
comes with battery, two wheels, caster,
motors, encoders, and a front ring of sonars.
The robot must have its microcontroller, as
well as a sonar board, power board,
ARCOS microcontroller server software, on
the hardware I/O bus with ARIA software
and ARIA Robotics API for software
developers (to add to the robots skills),
and an operators manual.
Geerhead.qxd 2/5/2008 5:58 PM Page 10
GEERHEAD
world to model everything the robot per-
ceives and responds to in the real world.
This way, the robots intelligence
can attempt to sort of reason out what
happens in the real world by use of the
stored working memory elements of
the virtual world, according to a Pace
Lab Obstacle Avoidance paper, by Dr.
D. Paul Benjamin, et al.
Machine Vision
The cognitive, machine vision robot
Pioneer 2 sees through a pair of
FireWire (IEEE 1394) Canon VC-C4 cam-
eras and framegrabbers. The cameras
are mobile via a pan/tilt hardware device
from Directed Perceptions. The computer
uses an onboard Linux OS for command
and control, which interfaces wirelessly
with a single PC-based computer.
The software brains of the Pioneer
2 consist of two parts: one that process-
es the bottom-up or simple view of
the world; and, one that processes the
top down or closer look view.
The bottom up view is quick and
dirty, giving the background and the
general lay of the land. This image
simply provides a view with objects the
robot may want to examine more
closely. This software runs on Intels
Open Vision software library.
The top down vision system is
much more complex and elaborate. The
robots system activates this element of
the robots vision when the robot wants
to take a closer look at something in the
environment that the first vision system
has only vaguely mapped out. This is for
specific object recognition.
This ERVision software will look
at an object to recognize its distinct
aspects. It will then model the image in
the virtual world and store that so the
object wont have to be recognized
again. This saves time and effort.
As the virtual world becomes more
completely detailed, the robots memo-
ry and perception of its world becomes
more clear and accurate. This way, the
robots obstacle avoidance proficiency
increases in its surroundings.
In case any of the obstacles should
be mobile objects (say a house cat, for
example, that might have been sleeping
in one spot when it was originally recog-
nized), the robot is able to determine
whether any objects it has recognized
before are now where they are sup-
posed to be (whether they have moved)
by use of the virtual world model.
SOAR
SOAR, an architecture for
developing intelligent robot systems,
has been in use for 25 years. It is
an integral part of the project that has
produced the Pioneer 2 robot (shown
in the images) with activevision.
The current version of the software
is SOAR v8.6, for those who would like
to investigate its use on their next
project. SOAR developers hope to bring
SOAR to the point where it can enable all
the tasks of the kinds of intelligent
robots the world envisions for the future.
The goal is for SOAR to solve open-
SERVO 03.2008 11
Camera and sonar view of Pioneer 2.
Pioneer 2 with camera mount and cabling, top
view. The P2 is the predecessor to the P3-AT
from ActivMedia. This model has four wheels
and motors with encoders. The AT model has
optional sensing software to make utilize
the new sensing hardware and turn that into
intelligible commands the robot can follow.
The robot comes with upgrades including
inertial correction that counters skid steer
dead reckoning errors. The bot can be
accessorized with laser range finders, both
front and rear sonars, pan/tilt/zoom color
cameras, stereo range finder cameras, and
day/night vision cameras. Finally, GPS,
color-tracking, compasses, and tilt position
sensing hardware round out some of
the most desireable add-ons. Internet
operation is also available with this model.
Another angle view of the Pioneer 2.
Pioneer 2 with a view of serial cable.
12 SERVO 03.2008
ended problems where, for example, there may not be a
single right answer. The developers want the architecture to
learn and use a variety of knowledge bases and problem solv-
ing skills. They want it to enable robots to interact intelligent-
ly with the world around them. They want to enable robots to
learn more about their own activities, tasks, and behaviors.
SOAR bases perception and action on all existing
knowledge including the latest interpretation of inputs from
the outside world. SOAR follows a robot AI model that
appears to closely resemble how human beings process
information for the purposes of perception and response.
Working forward from version 8, developers are seeking to
store multiple representations and interpretations of acquired
knowledge, as well as to acquire that knowledge in different
ways. SOAR brings all this knowledge to bear on every
perception, decision, and reaction at the softwares runtime.
Developers are turning SOAR into a suite of cognitive capa-
bilities matching those of the human brain. SOAR can retrieve
knowledge and memories of previous perceptions and reactions
and model those to determine how to react in the current state.
Fine Tuning
In order to make sure the virtual world as perceived and
modeled by the robot and the real world match up, the
robots rules test for differences between the two. If there is
a new object, if an object has moved or changed, there are
rules to deal with these differences.
To do this, SOAR and the vision system work to collect
the more vague bottom up information about segments of
visual data, stereo information (there are two cameras), and
motion information.
The SOAR system applies rules to the robot to force it to
examine the object from the top down by turning its stereovision
to focus on the object more exactly. System rules process the top
down vision data by examining portions
of the object via the object recognition
software. The results of those examina-
tions are stored in SOAR. SV
GEERHEAD
Pace Robotics Lab
http://csis.pace.edu/robotlab
Pace robot clips including
the Pioneer 2
http://csis.pace.edu/robotlab/
movies.html
Pace robot clip
clips/RobotMovie1.mov
Pace robot clip
clips/RobotWorldModel.mpg
Large Pace robot clip
(save file before viewing)
clips/puma.avi
RESOURCES
P
erform proportional speed, direction, and steering with
only two Radio/Control channels for vehicles using two
separate brush-type electric motors mounted right and left
with our mixing RDFR dual speed control. Used in many
successful competitive robots. Single joystick operation: up
goes straight ahead, down is reverse. Pure right or left twirls
vehicle as motors turn opposite directions. In between stick
positions completely proportional. Plugs in like a servo to
your Futaba, JR, Hitec, or similar radio. Compatible with gyro
steering stabilization. Various volt and amp sizes available.
The RDFR47E 55V 75A per motor unit pictured above.
www.vantec.com
STEER WINNING ROBOTS
WITHOUT SERVOS!
Order at
(888) 929-5055
The International Business Development Event for the
Personal, Service and Mobile Robotics Industry
April 8-10, 2008
Pittsburgh, PA
David L. Lawrence Convention Center
BUILDINGAN INDUSTRY
Now in its fth year RoboBusiness Conference & Exposition is the must-attend event for those
interested in the business and technical issues related to the development of the personal, service and
mobile robotics industry.
WORLD CLASS SPEAKERS
RoboBusiness 2008 features two days of keynotes & general sessions delivered by internationally
recognized leaders from business, government and academia. RoboBusiness tracks include:
Q Business Development and Partnership
Q Applications and Products
Q Achieving Autonomy New!
Q Workforce Development New!
Q RoboMedicus/Healthcare Robotics New!
Q Security and Defense
Q Intelligent Transportation and Field Robotics New!
Q First Responder Robotics New!
Q Consumer Robotics
EXPOSITIONAND SPECIAL EVENTS
Q Expo oor featuring over 75 companies from around
the world
Q Opening night Networking Reception
Q Open Door Keynote New for 2008
iRobot CEO Colin Angles keynote open to all attendees
Q Robot Hall of Fame reception and induction
ceremony New for 2008
Q Site visits to Carnegie Mellon Robotics Institute and
National Robotics Engineering Center New for 2008
For more information or to register, visit
www.robobusiness.com or call 800-305-0634.
For sponsorship opportunities, contact
Ellen Cotton at 508-663-1500 x240.
www.robobusiness.com
Register by
March 14th and
Save $200 on a Full
Conference Pass
Register by
March 14th and
Save $200 on a Full
Conference Pass
Founding Sponsor
Premier Sponsor
Gold Sponsors
Corporate Sponsors
Analysts, Associates
& Academic Co-Sponsors
Media Co-Sponsors
Produced by
Full Page.qxd 2/4/2008 1:36 PM Page 13
14 SERVO 03.2008
s a mechanical engineering
student, Evan has been learning
about things like how Bessel
functions are the eigenfunctions of the
Sturm-Liouville Equations that can be
used to describe heat conduction in
nonrectangular geometries, and how
computational methods like the
Newton-Raphson method can be used
to find the solution to large nonlinear
systems. Such highbrow concepts in
engineering however interesting they
might sound can only be mastered
with a firm grasp on the fundamentals
of engineering and physics.
In our last article with the V-Bot,
we alluded to the simple R/C cars that
we outfitted with aluminum armor and
weapons for rumbles. While these
simple emulations of the action we had
seen on Battlebots might have seemed
like nothing but wholesome fun, we
were actually learning the fundamen-
tals of engineering that would be the
foundation of our success in big bot
competitions like FIRST and beyond.
Keeping the Team
Together
Our first foray into robotics came
when we were in middle school, and
with the help of our dad and
scavenged parts from Cosworth
Racing, we competed in Botbash 2001
with our 60 lb entry Troublemaker. The
thrill of competition inspired us to
spread the joy of robotics to our peers,
and when we got to high school we
formed a robotics club, Club CREATE
(Chaparral Robotic Engineers and
Techno Explorers).
Our original ambition with Club
CREATE was to enter a combat robotics
competition, but the design and
fundraising process was daunting and
slower than the lowest setting on a
crock pot. Brainstorming sessions
began to lose their luster when we
didnt have much to actually work on,
so we had to think of something else
to do; something fun, yet productive
and instructive.
To keep the team interested and to
hone our skills in combat driving, we
organized a series of rumbles with
souped-up R/C cars. The extent of our
modifications was basically to outfit
our cheap R/C cars with aluminum
weapons, but that was already a good
lesson in some shop skills. We learned
to use aviation shears, hacksaws, and
jigsaws, and while these skills may be
taken for granted by many folks that
have been hobbyists or professionals
for a long time, you have to learn
sometime.
The R/C car rumbles were a great
project to practice our shop skills on,
because it was just for fun and conserv-
ing time and material was not an issue
like it would be in a competition like
FIRST. So we were able to build our
confidence with tools in a relaxed
setting, and with all the metalworking
we did, we even became acquainted
with the properties of materials.
The Burning Means
Its Working
Our most sophisticated endeavor
in crafting our weapons was a
rudimentary foray into heat treatment
of materials. We employed used
engine oil as our heat treatment fluid,
and the objects of our treatment were
nails that we thought lacked the bite to
be truly fearsome. We used an
acetylene torch to heat up the future
instruments of destruction, and then
we quenched them to seal the deal.
While we might not have known the
subtleties of the effects of heat
treatment and quenching on grain size,
simply becoming familiar with the
THIS MONTH:
Back to Basics
TwinTweaks.qxd 2/5/2008 5:53 PM Page 14
process was still a valuable lesson.
Our tutelage in materials science
included a number of other topics that
may have been somewhat less exciting,
but also undeniably useful. Working
with our makeshift aluminum weapons
was a good introduction to the
different alloys of aluminum. We
learned useful tidbits about how 7075
aluminum is unweldable and will tend
to fracture instead of bend, and
that 6061 aluminum is much
more amenable to being bent and cut
by aviation shears.
We picked up other useful tidbits
like how the number at the end of the
alloy designation (like T6) referred to
the heat temper. And again, even
without knowing the subtleties of the
International Alloy Designation System,
simply being able to identify the
different alloys was a great first step.
Our endeavors to create the most
fearsome weapons also introduced us
to a variety of materials, most notably
titanium, which loomed imperiously in
our untrained minds as something
particularly formidable. Like all
materials, titanium has its uses and its
drawbacks, but we simply saw our
titanium valves as the perfect way to
give our combatants a sophisticated
bite. They may not have been
particularly sharp, but they were
titanium, so the valves were cool.
And in our situation where
scavenged materials were plentiful,
discussions about where they came
from (in this case, racecar engines)
were natural and informative.
Not all of our lessons in
materials came with the cold luster
of assorted metals we were
also introduced to the world of
reinforcing materials like carbon
fiber. We appreciated the carbon
fiber for its cool factor and its good
strength-to-weight ratio, and even
though we werent referring to
things like ultimate tensile strength,
resilience, and compressive strength,
the project was still helping us to
develop that intuitive sense of what
material is effective for a certain
application. Such an intuitive sense is
certainly useful for many projects
during the initial brainstorming stage,
and it certainly buoys confidence in
ones problem solving abilities.
And, of course, we became well
versed with that most indispensable of
all the engineers favorite materials
duct tape. Competitors like the Duct
Tape Avenger can attest to the univer-
sal applicability of the stuff, even if it is
not the most elegant of solutions
(though not for lack of trying).
Other simple lessons were also
instilled in our budding scientific
minds during the building phase,
like everything from the virtues of
conserving material (try cutting that
piece from the edge instead of the
center so you can get another spike
out of it) to the benefits of putting
bevels in plates to improve strength.
Just learning the little tricks that
experience teaches was a great
benefit of the project.
Are You Ready to
Rumble?
To turn our cheap R/C cars into
fearsome competitors, we had to
consider more than just weapons the
scariest weapons would be useless if
they were on a robot that couldnt stay
upright. One of the important lessons
we learned in the rumbles that applies
to everything from R/C cars to FIRST
robots was the importance of keeping
track of the center of gravity.
Tall robots might look intimidating
while upright in their full glory, but
they become much less terrifying once
they have inevitably tipped over.
Bryces low profile Quagulis was nearly
impossible to flip over and, as a result,
was one of the top competitors in the
rumbles. Evans Duct Tape Avenger
NIDHOGG.
MOINSPIRE.
MOCK-UPS OF QUAGULIS (LEFT) AND DUCT TAPE AVENGER (RIGHT). QUAGULIS.
SERVO 03.2008 15
Back to Basics
16 SERVO 03.2008
Twin Tweaks ...
was a higher profile robot that often
found itself on its side after sustaining
a well placed hit at the hand of one of
its rivals.
Center of gravity may seem like
another one of those intuitive con-
cepts, but crashing R/C cars are much
more exciting display of physics than,
say, the equation for center of mass:
R = (r) r dV
(r) dV
which basically integrates a position
weighted mass density and divides that
by the total mass. This formula is much
less intuitive to the untrained mind,
and this points to another benefit of
the rumbles all of the lessons we
were learning would eventually be for-
malized by formulas and theories inside
the classroom and lecture hall, but our
adventures with Quagulis and Nidhogg
gave us a personal context that
allowed us to visualize and learn all of
these lessons more easily.
Keeping Our
Momentum
The rumbles were also very
instructive on the subject of
momentum. While smaller competitors
like Miscreant and the Masking Tape
Avenger might look fierce, they were
tragically outmatched by the sheer
size of bots like Slag. The essence of
this disadvantageous match-up was in
the difference in momentum between
the competitors. Momentum is given
by the equation:
p = m*v
where p is momentum, m is mass, and
v is velocity. Some of the smaller cars
did have comparable speeds with the
bigger competitors, but they were sore-
ly outmatched in the mass department.
This points out the benefit of separat-
ing robots into different weight classes,
which we did in our rumbles. In the
realm of combat robots in particular,
change in momentum can be related
to force by the impulse equation:
F*t = m*v
where F is force, t is the time
interval over which the force is
applied, m is the mass of the moving
object, and v is the change in
velocity due to the impact. Many
prospective engineers have an
intuitive sense of this phenomenon,
but seeing it put into practice as one
armored R/C car runs into another
with spikes clashing and tires
squealing is enough to spark the
curiosity to ask the question as to why
it actually happens in the first place.
Once again, this is a time when we
were developing a context for our later
formal training, and all the talk about
elastic and inelastic collisions and
conservation of momentum could be
associated with exciting images and
memories in addition to the equations
and pictures in the textbooks.
Sporque Not Just
a Freakish Utensil
Anymore
Occasionally, the rumbles could
involve some pushing matches and
other displays of brute force, and such
incidents were very enlightening in the
area of speed and torque. Quicker
bots like the Duct Tape Avenger might
look intimidating when they reached
ramming speed, but slower, torquier
bots like Slag would inevitably win in a
shoving match.
Of course, these comparisons are
not completely fair because the R/C
cars may have been graced with differ-
ent motors, but it was still certainly
illustrative of the unavoidable tradeoff
between torque and speed. In
an ideal world, it would be
wonderful to have both high
torque and high speed, but the
rumbles did us the service of
opening our eyes to the harsh
reality that you cant always get
what you want.
A tentative foray into
motorized weapons was able to
demonstrate an extreme case
of the torque/speed tradeoff.
We were able to get our hands
on some nondescript high
speed motors, and they seemed
SLAG.
MASKING TAPE AVENGER.
MISCREANT.
to be the perfect way to create some
menacing spinning weapons. After
outfitting the shafts with some cut
aluminum, attaching an extra power
source of a couple AA batteries, and
fastening the whole thing together
with generous amounts of duct tape,
we couldnt deny that the spinners
looked positively ferocious on little
bots like Nidhogg. The aluminum
blades spun at such speed that they
became a blur, and to look at them
one would likely think that they
were capable of inflicting some
serious damage.
As those acquainted with the laws
of physics might have already guessed,
the spinning weapons were, to say the
least, disappointing. They would come
to a screaming halt after what seemed
like a simple love tap on its rival, and
the worst damage they could do was
purely cosmetic. And on top of all that,
they even had a slightly noticeable
effect on the movement of the R/C car
that was sporting it turning became
more difficult. We may not have
built an effective weapon, but we did
unintentionally create a passable
gyroscope.
Despite the ineffectiveness of the
spinning weapons, the whole scenario
was the perfect segue into a discussion
about gear ratios. Many times,
discussions that make ample use of the
terms ratio, radius, and angular
velocity might cause an uncooperative
students eyes to glaze over in a bored
stupor, but the rumbles once again
gave an exciting context to the
discussion. Talks like these werent just
about geometry and enigmatic Greek
letters, they were about improving the
weapons on your bot so that next time
you might just be able to give
Miscreant whats coming to it.
C. F. E.
The rumbles also began our life-
long education in the subject of Cool
Factor Engineering. Making some-
thing functional may be paramount,
but we think the penultimate goal of
any project should be to make
something that you are proud of. In
general, we find that we are proud of
a project when it is interesting, eye
catching, and simply exudes an air of
coolness. In the case of our R/C cars,
that might mean adding plates of
carbon fiber, or using titanium vales
for weapons instead of cut aluminum.
The tiny competitor Bucephalus was
a perfect example of how to
achieve cool factor through a careful
combination of sculpted aluminum
and Greek mythology.
In later projects like FIRST robots,
cool factor might be achieved with
clean wiring, smooth bottom panels by
virtue of countersunk screws, or
colorful stickers that gave credit to our
generous sponsors. Whatever form it
might take, cool factor is one way that
we and our teammates have been able
to establish that all important
ownership of our projects. Personal
touches and details meant to add just a
little bit of flair are important, because
they can be the deciding factor
between pointing to a project and
saying I made that and it completes a
task versus I made that, it completes
a task, and its awesome.
Cool factor also plays a role in the
larger world of robotics, and especially
in the arena of public opinion. Case in
point: The Neiman Marcus 2007
holiday windows promised robots
decorating a Christmas tree, and when
the curtains were pulled back to reveal
automotive assembly line robots
dutifully placing ornaments, some
expectant onlookers asked themselves
where are the robots? But when peo-
ple see Hondas Asimo on television
seemingly leading the march to the
future, they are hard pressed to
identify it as anything other than a
robot. Automotive assembly robots,
though undeniably functional, lack the
cool factor required to make them
universally recognizable as robots.
Folks inside the technology indus-
tries may have no problem identifying
the indefatigable arms as fine
specimens of the robotic species, but
many people are looking for something
that looks like Rosie from the Jetsons
or a Transformer to fit their definition
of a robot. Critics of this perception of
robots may argue that designs that
look like Optimus Prime are not nearly
as effective at vacuuming floors as
a Roomba, but they probably also
wouldnt deny that it would be totally
awesome to have Megatron doing
your dishes.
Some research scientists might
eschew public opinion as inconsequen-
tial for robots that are firmly rooted in
R&D laboratories, but commercial
robotics companies need to appeal to
their consumers through sleek designs
that are undeniably benefited by the
addition of a bit of cool factor.
Cool Factor Engineering fosters a
sense of pride in ones work, and
shows that functionality and aesthetic
value are not mutually exclusive. Its
enough to challenge Oscar Wildes
adage that all art is quite useless.
Fun: Part of a Balanced
Intellectual Diet
What were trying to say is that
even a really simple project that seems
more concerned with fun than with
hard scientific principles can still
reinforce a solid foundation in
engineering. Even if the participants
dont know the exact terminology for
center of gravity or tank style drive
train, they are still learning the
principles, and they are developing
that all important intuitive sense of
what works and what doesnt.
Simple projects like these are also
Back to Basics
BUCEPHALUS.
SERVO 03.2008 17
a great way for parents to mentor their
kids, even if they dont have the
technical background of the mentors
on FIRST teams that know the ins and
outs of C programming and control
theory. All an effective mentor needs
to do is start asking the right questions
why do you think the Duct Tape
Avenger flipped over? or how do you
think you can make that flamberge-
esque spike stronger?
Kids can learn to celebrate the
search for answers, and according to
Dean Kamen, we get what we
celebrate. Our R/C car rumbles like
their larger combat robot inspirations
dont celebrate violence or destruc-
tion; they celebrate the competitive
testing of ideas, and the prospect of
customizing an R/C car for battle like
bigger bots on TV might be just
enough to pull Junior away from the
Guitar Hero for an afternoon. Just
instilling that positive association with
problem solving and science is the
first step to helping them become
successful engineers, and thats
something any mentor can do no
matter what their level of technical
expertise.
Little fun projects like our R/C car
rumbles can also instill that sense of
fun and excitement that kids dont
learn to associate with engineering
from doing free body diagrams and
repetitive calculus problems. It can
expose them to the exhilaration of
problem solving, even if the problem is
finding the way to best inflict mortal
damage upon your mechanical
opponents. And our R/C car rumbles
really did that, by keeping the robot
club together and giving us the
foundation we needed to succeed in
competitions like FIRST and beyond.
The thrill of competition and the
excitement of duels between rivals like
Quagulis and the Duct Tape Avenger
kept the imaginations of the team up
and running many times the small
scale tussles would inspire ideas that
we planned to put into practice on a
bigger robot.
At first, our target big bot was the
proposed combat robot, but upon
discovering FIRST and the Kleiner
Perkins Caufield and Byers grant, we
decided to change course and follow
the path of least financial resistance.
And, in truth, if you look at our robot
MO, built for the 2003 FIRST Season
(Stack Attack), you can see that many
of the lessons and inspirations of the
rumbles finally found their form. MO
had a low center of gravity, slanted
sides reminiscent of combat robot
wedges, and it was decked out
with the very same aluminum alloys
that we used when fashioning our
pint-sized weapons.
After a successful run in competi-
tions ranging from FIRST to PAReX
(Phoenix Area Robotics Experimenters)
to the Solar Cup, the team members
that participated in those rumbles are
now sprinkled throughout the country
as engineering students at top notch
universities, and theyre living proof
that fun does a brain good. SV
Twin Tweaks ...
18 SERVO 03.2008
Whats the difference?
Price!
Youre correct, the one on the right costs
less because its from Jameco! When
youre looking to purchase major name
brand passives or semiconductors, be
sure to check out Jameco first. They also
offer money-saving generic equivalents
at about 20% less, so youll save even
more! And if you should find a lower
print-advertised price for a product
Jameco offers, just let them know and
theyll offer that product for 10% less
than the other companys price, right
down to Jamecos cost! Wouldnt it be
worth your time to check out Jameco
today at www.Jameco.com?
Both Major Name Brands and Generic Products for Additional Savings
Over 99% of Catalog Products are in Stock for Immediate Delivery
Only Jameco Guarantees the Lowest Prices
Click or call today for a FREE Jameco Catalog
& start benefiting from Jamecos Awesome Prices!
Visit www.Jameco.com/NVT
Call 1-800-831-4242
Know of any robot competitions Ive missed? Is your
local school or robot group planning a contest? Send an
email to steve@ncc.com and tell me about it. Be sure to
include the date and location of your contest. If you have a
website with contest info, send along the URL as well, so we
can tell everyone else about it.
For last-minute updates and changes, you can always
find the most recent version of the Robot Competition FAQ
at Robots.net: http://robots.net/rcfaq.html
R. Steven Rainwater
M Ma ar r c ch h
7-8 AMD Jerry Sanders Creative Design Contest
University of Illinois at Urbana-Champaign, IL
Check the website for the details of this years contest.
http://dc.cen.uiuc.edu
7-8 National Robotics Challenge
Marion, OH
Offers a variety of robot events for middle school,
high school, and post-secondary students. In
addition to traditional robot Sumo, students can
also create models of an industrial robot work cell
or the most interesting build robots that can
create graphic works of art.
www.nationalroboticschallenge.org
8 Fort Collins Robot Fire Fighting Challenge
Discovery Science Center, Fort Collins, CO
This is a regional for the Trinity College Fire Fighting
Robot contest. Autonomous robots must locate
and extinguish a flame in a scale model of a home.
www.strout.net/fcrf fc
15-16 Manitoba Robot Games
Tec Voc High School, Winnipeg, Manitoba, Canada
Included in this competition are a mix of events
for autonomous and remote-controlled robots
including Japanese style mini-Sumo, Western style
Sumo, a robot Mini-Tractor Pull, Super Scramble,
line-following, and the Robo-Critters contest for kids.
www.scmb.mb.ca
15-16 Roboticon
University of Guelph, Ontario, Canada
The robot contest is part of a larger University
Open House day which includes contests in every-
thing from flower arranging to pancake flipping.
www.collegeroyal.uoguelph.ca
29 CIRC Central Illinois Bot Brawl
Lakeview Museum, Peoria, IL
This event includes RC combat, autonomous
Sumo, line-following, line maze. Autonomous and
remote-control robots.
http://circ.mtco.com
30 Boonshoft Museum Robot Rumble
Boonshoft Museum, Dayton, OH
This event includes robot building and competition.
www.boonshoftmuseum.org
TBA Penn State Abington Mini Grand Challenge
Penn State Abington, Abington, PA
Includes outdoor autonomous mobile robot navigation.
www.ecsel.psu.edu/~avanzato/robots/
contests/outdoor/contest05.htm
TBA DPRG RoboRama
Dallas, TX
Includes quick trip: an entry-level event where the
robot moves from A to B and back; line-following:
the robot must follow a line on the floor; T-time:
the robot moves through three points on a
T-shaped course; and can can: the robot must
locate and retrieve empty soda cans.
www.dprg.org/competitions
A Ap pr ri i l l
26 RoboFest
Lawrence Technological University, Southfield, MI
Includes game competition two autonomous robots
work together. Also robot exhibition, RoboSumo,
RoboFashion show, and mini urban robot challenge.
http://robofest.net
12-13 Trinity College Fire Fighting Home Robot contest
Trinity College, Hartford, CT
The well-known championship event for fire
fighting robots.
www.trincoll.edu/events/robot
Send updates, new listings, corrections, complaints, and suggestions to: steve@ncc.com or FAX 972-404-0269
SERVO 03.2008 19
Events.qxd 2/6/2008 10:26 AM Page 19
20 SERVO 03.2008
Q
. I would like to know your opinion on performing
some simple modifications to my SumoBot from
Parallax so that it can be used to solve line mazes. It
is my understanding that I would need some encoders on the
wheels so that it can keep track of how far it has moved so
that it can repeat its previous paths. I am not an electronics
expert, so I am hoping that you can point me to something
that is more plug-and-play than build from scratch. Any help
would be greatly appreciated.
Lynn Brown
A
. This sounds like a fun project! I think I have the ideal
set of plug-and-play products for you. Yes, encoders
can be a big help in solving a line maze, but they are
not required. Many people do this without encoders.
As for simple off-the-shelf, plug-and-play encoders that
will work great with the SumoBot, take a look at the
WheelWatcher encoders from Nubotics (www.nubot
ics.com). These encoders are designed to mount directly
to regular servo motors (that have been modified for
continuous rotation) like those used with the SumoBot.
These encoders will keep track of how many degrees each
wheel rotates and the direction of each rotation. These
encoders also come with a self-adhesive encoder disk that is
mounted to the inside of the wheel. With the SumoBots 2.6
inch diameter wheels, these encoders by themselves can
keep track of positional accuracy of at least 0.25 inches of
travel ( * 2.6 inches/32 counts per revolution = 0.25 inches
per count). Technically, all the BASIC Stamp has to do is keep
track of the encoder counts for each wheel as the robot
moves through the maze.
But, having the BASIC Stamp keep track of all of
the encoder counting, direction
control, turning, updating the
servo motion commands every 20
ms, and monitoring the line
sensors while trying to solve the
maze is going to be a bit of a
programming challenge. Thats not
to say it cant be done since many
people have been very successful
in doing this, but there is another
plug-and-play device that will
greatly simplify all this.
The WheelCommander from
Nubotics (Figure 1) is a closed loop
differential servo motor controller.
With this controller, all you have to
do is tell the robot how far you
want it to go, how fast you want it
Tap into the sum of all human knowledge and get your questions answered here!
From software algorithms to material selection, Mr. Roboto strives to meet you
where you are and what more would you expect from a complex service droid?
by
Pete Miles
Our resident expert on all things
robotic is merely an Email away.
roboto@servomagazine.com
Figure 1. The WheelCommander closed loop
differential servo drive motor controller.
Figure 2. Original Parallax SumoBot
before modifications.
Line Maze is typically a contest where a line is placed down
the center of a maze puzzle. There is no wall in this type of a
maze and the robot must use the line to solve the maze. Many
contests allow the robot to run through the maze several times,
and the fastest run time is used for the final score. Remembering
the maze path helps to greatly reduce the amount of time
required to solve the maze on subsequent attempts. One
source for a complete set of rules for this type of a contest can
be found at the Robothon website (www.robothon.org).
LINE MAZE
MrRoboto.qxd 2/5/2008 5:40 PM Page 20
to move, or what angles
you want it to turn (there
are many other parame-
ters that you can control).
You no longer have to
worry about controlling
and monitoring each
individual wheel to do all
of this. The robots motion
is now controlled by a
small set of RS-232 serial
commands or I
2
C com-
mands, and best of all, the
BASIC Stamp no longer
needs to update the servo
positions every 20 ms. This
frees up the BASIC Stamp
to focus on monitoring the line sensors and solving the maze.
When the robot reaches an intersection, it can query the
WheelCommander to see how far it has moved and can
record this value for later analysis.
Figures 2 through 13 show some assembly steps in
upgrading the Parallax SumoBot with WheelWatcher
encoders and the WheelWatcher Commander for motion
control. Figure 2 shows the original Parallax SumoBot (before
modifications) with the opponent infrared sensors removed.
Figure 3 shows the jumbled mess of wires after the BASIC
Stamp embedded in the SumoBoard has been removed.
Figure 4 shows the servo removed from the SumoBot
next to all of the parts that come with a WheelWatcher kit.
All of the nylon spacers and washers shown here are not
used in the assembly process. Depending on the geometry of
the servo motor, different combinations of washers and
spacers are used to ensure that the WheelWatcher board is
properly spaced on the servo.
The manual that comes with the WheelWatcher lists
several spacer and washer combinations to use with several
different types of servos. The Parallax servo shown here was
not included in the list in the manual, but it uses the same set
of spacers as the Futaba S3001; so use the short spacers and
the thick washers in the assembly process.
Figure 5 shows the WheelWatcher mounted on the
servo. The clear plastic disk on the servos output spline is the
alignment tool that comes with each WheelWatcher kit. This
tool ensures that the encoders are mounted with the proper
orientation with respect to the servo shaft. Figure 6
shows the servo and WheelWatcher mounted back on the
Parallax SumoBot.
The existing Sumo ring edge sensor cannot remain
mounted in its original position using the existing 1-1/4 inch
long aluminum spacer. This will interfere with the connector
on the WheelWatcher. It will have to be removed. Since line
sensors on line following robots are located more towards
the centerline of the robot, I turned the sensors around and
used the two existing holes at the base of the robot frame as
new mounting locations.
To conserve parts, I used two of the long spacers from
the WheelWatcher kit and the small nylon spacers that were
on the original sensor post with a new 4-40 x 5/8 long
screw to remount the sensors. Figure 7 illustrates these
components and Figure 8 shows the bottom of the robot
with the new sensor orientation. Note that the distance
between the sensors is now about one inch.
SERVO 03.2008 21
Figure 5. WheelWatcher mounted on the servo. Figure 6. WheelWatcher and servo mounted
back on the Parallax SumoBot.
Figure 4. Servo, wheel, and WheelWatcher components.
Figure 3. SumoBoard removed
from the SumoBot.
The best place to mount the WheelCommander is right
on top of the servos (directly under the SumoBoard). Due to
vibrations in the robot and flexure between the two servos, I
decided that I was going to place a small board between the
WheelCommander and the servos. Figure 9 shows some
double sided foam tape added to the servos for mounting
the intermediate board, and Figure 10 shows a 1/4 thick
piece of scrap Sintra (expanded PVC). You can use different
types and sizes of material here. This was just something I
had lying around. I then used two more pieces of double
sided foam tape to attach the WheelCommander to this
board (see Figure 11).
Figure 12 shows all
the wiring attached to
the WheelCommander. You
should notice that there is a
new battery attached to
the front of the robot. The
WheelCommander requires
a minimum of 6.2 volts to
operate properly. Since the
regular SumoBot uses
only four AA batteries
(6.0V), they were insuffi-
cient to power the
WheelCommander along with the BASIC Stamp and drive
the servo. So, I added a 9V battery to the front of the Parallax
SumoBot. This battery fitted nicely in front of the servos and
WheelCommander. A rubber band was used to hold the
battery to the screws that hold the servos in position.
Figure 13 shows the completed robot. The servo power
to the WheelCommander is plugged into the B and R pins
used to power the servos on the original SumoBoard. I chose
to do it this way so that I can take advantage of the
three-way power switch on the SumoBoard to make sure
that the servos are not accidentally powered during
rogramming efforts. The rest
of the wiring is pretty much
the same as outlined in the
WheelCommander manual.
Ideally, you would add a
power switch to turn the power
on to the WheelCommander. I
chose to use simple connectors.
There are two things you
are going to want to download
from the Nubotics website.
The WC Wizard program
and the example program
(wc116_bs2_demo.bs2) for the
BASIC Stamp. The WC Wizard is
a great utility for testing and
configuring the WheelCommander
for your specific robots
geometry. Also with the WC
Wizard, you can tune the PID
(Proportional, Integral and
Derivative) constants for the
closed loop position and velocity
control of your robot.
In addition, you can change
the serial communication baud
rate for the WheelCommander.
The default baud rate is 38400.
Now all of these configuration
parameters can be transmitted
to the WheelCommander via
the BASIC Stamp. The WC
Wizard has some nice visual
22 SERVO 03.2008
Figure 7. Reusing parts for sensor relocation. Figure 8. New edge/line sensor orientation.
Figure 11. WheelCommander mounted
on the servos.
Figure 12. All wiring attached to
the WheelCommander.
Figure 9. Double sided foam tape for
mounting the WheelCommander to the servos.
Figure 10. Intermediate board between the
servos and the WheelCommander.
feedback to help with the PID tuning.
To avoid a huge problem I had with using the WC
Wizard, make sure that the RS-232 adapter between your
computer and the WheelCommander converts the signal to
standard TTL voltage levels (0-5V) and that it inverts the l
ogic levels, or the two devices wont communicate with
each other. Also, make sure that the CTS and RTS lines
are jumpered together. The WheelCommander manual
recommends several adapters from Acroname (www.
acroname.com) who is also a distributor of the Nubotics
products.
It is highly recommended that the baud rate be changed
from 38400 to 9600 for the BASIC Stamp that is embedded
on the SumoBoard. This is because at higher baud rates, they
dont synchronize together properly.
In the example program, there is a subroutine called
initfw (see below). The call to this routine is normally
commented out since it only needs to be called once. The
first command F0302 changes the baud rate to 9600. If
the baud rate was changed by the WC Wizard, this line can
be commented out. The command F0289 MUST be execut-
ed one time successfully after the baud rate has been
changed to 9600. Otherwise, the BASIC Stamp wont be able
to effectively communicate with the WheelCommander
(actually setting bit 0 to high for the Mode Constant is
needed). This command adds a critical time delay in the seri-
al communication strings that is needed for the BASIC Stamp
to properly synchronize and transmit all bytes back and forth.
initfw:
SEROUT 1, 6, [F0302, LF] at 38400, set baud = 9600
GOSUB getack
SEROUT 1, 84, [F0289, LF] mode = slow comms
GOSUB getack
RETURN
At this point, you should be ready to start programming
your robot to solve line mazes. Like I said before, this sounds
like a fun project. When you get your robot up and running,
write a short article about what you did and submit it to
SERVO Magazine. I am sure many readers would love to
learn and see what you did.
Q
. I have noticed over the years that you use the
BASIC Stamps and the SX microcontrollers in most
of your examples. Why is that, and why havent you
talked about the new Propeller microcontroller from Parallax?
This looks like a very interesting microcontroller, especially
with its multitasking capabilities.
Shawn Kidwell
Montgomery, CA
A
. I suppose that some people may think I am biased
towards Parallax (www.parallax.com) products. Well,
to tell you the truth, I am biased towards them.
Especially in the context for this magazine and the way I write
my articles. The way I look at all of this is that most of the
people that read this magazine want to learn how to do
something. I have chosen a writing style that tries to explain
how you would go about solving various challenges with real
working examples, along with explanations of how things
work and also showing some pitfalls.
This teaching style is one of the main reasons I like work-
ing with Parallax products. They have the best documenta-
tion in the world on how to use their products (along with
many other topics such as basic electronics) with many prac-
tical examples. Their documentation style is about teaching
people how to do things from scratch, taking you from little
to no experience to making you a competent embedded
microcontroller applications designer/robot builder. If you
run into a problem, give them a call or go to their online
forums, and they will bend over backwards to help you.
If after building one of my example products, you want
to learn more about what you can do with a BASIC Stamp or
with the SX microcontroller or even with the Propeller
you can find the answers on the Parallax website or on their
forum pages (http://forums.parallax.com).
I am not saying that other microcontrollers and
documentation is bad. They have their places, and they have
some very devoted supporters/developers. In most applica-
tions, just about any microcontroller will work just as well as
any other microcontroller. Some just do certain things better
than others. That is why there are so many to choose from.
In most applications, Parallax products will do a fine job.
As for the Propeller chips from Parallax, to tell you the
truth, I havent really had the chance to dive into them until
recently. And now that I have, I wish they were around
many years ago when I first got involved with robotics. For
robotics applications, the Propeller chips are probably the
best microcontroller out there. The reason I say this is
because it can do multiple things at the same time. With
other microcontrollers, we spend a lot of time trying to
figure out how to write code that can continuously monitor
its environment, make decisions on what the sensors are
telling it, and controlling all the functions of the robot.
Trying to get the right timing of all of these different things
SERVO 03.2008 23
Figure 13. Completed reconfiguration of the Parallax SumoBot.
often becomes very difficult, and in many cases, desired
capabilities are scrapped because proper timing cannot
be executed.
Here is a different way to think about robotic
programming: Habits. Yes, habits. Lets use an interesting
illustration driving a car. When you first drove a car (for
those of you that are not old enough to drive, you will expe-
rience all of this), it was a very complex endeavor. Working
the gas pedal, brakes, clutch pedal, working the stick shift,
the steering wheel, turn signal, looking at and reading road
signs, watching out for all the cars in front of you, to the
sides, the rear, and planning your route to your destination.
These are many different things that are done at the same
time. But after driving the car for a while, you no longer had
to think about all of these things. Instead, you get in the car,
turn it on, and proceed to your destination. Now the only
thing on your mind is the next robot project you are
working on. All of the early trials in learning how to drive are
now habits that are automatically happening without any
conscious thought.
But when we program our robots, these basic little
habits are still a main part of the thought process. Thus, the
main microcontroller is spending all its time working on all
the tiny little details. Can you imagine how tough it will be to
drive, if every time you got behind the wheel it was like the
first time, and you have to actively process every little detail?
Traffic would definitely be lighter since most people would
give up and take the bus.
Now if we can off-load many of these little habits to
parallel processors, they can then spend all their time focused
on dedicated activities such as: infrared sen-
sor arrays with ultrasonic sensors for obstacle
detection; a video camera for tracking a red
colored object; PID motor speed controller; RF
data uploading and downloading; etc. All of
these tasks then talk to the main processor to
determine what to do next based on the data
from the various inputs. Each one would run
independently without having to deal with the
timing of the other processes. You could use
multiple microcontrollers that are dedicated to each task to
do the same thing. But every processor can share data
between each other on an as-needed basis without a main
processor having to coordinate the efforts between
them. This can open up a bunch of new programming/robot
behavior capabilities.
Take, for example, a remote control humanoid robot
using an off-the-shelf Playstation 2 wireless controller. I have
written several articles on how to simplify the communication
between devices using these types of controllers since the
actual data transmitted wirelessly is rather complex, and it
takes a certain amount of time to process. Now when a
humanoid robot is moving 17 different servos at one time,
there is a lot of control algorithms running to synchronize the
servos. If the human operator decided to tell the robot to
turn to the right while the robot is walking forward, the main
microcontroller will finish the current set of motions then
look for the remote control. It will send out a command, I
am ready, and the controller will respond with its current
state. Then the robot will respond to it. In many cases, by the
time the robot is ready to make that right turn, it is too
late, especially in a ROBO-ONE style of competition
(www.robo-one.com).
Now if a Propeller was being used to control the robot,
it could be monitoring the Playstation 2 controller
continuously, and when a new motion command is executed,
the robot will have continuous information about controller
status and can immediately respond without any lag
times between motion sequences. Believe me, the lag time is
rather annoying.
The Propeller can run up to eight different 32 bit tasks at
the same time. This is a ground up design from Parallax that
no other company has done a true multitasking microcon-
troller. With a clock speed up to 80 MHz and a total of 64
kbytes of memory, some very amazing things can be done
through the 32 I/O pins which can sink/source up to 40 mA
each. The Propeller has two different programming
languages. One is called SPIN and the other is assembly,
where SPIN is Parallaxs higher level programming language.
Figure 14 shows a photo of the Parallax Propeller demo
board with the H48C three-axis accelerometer module with
a NTSC LCD video monitor graphically displaying the
orientation of the accelerometer. All of these components
are available at Parallax and the source code can be
downloaded from their website. The Propeller is directly
controlling the video display.
One of the coolest features about this microcontroller
is that its programming environment is an object oriented
24 SERVO 03.2008
LCD Drivers Xbee Transceiver CMU Camera Tracking
TV Terminal Gamecube Controller GPS Drivers
VGA Drivers Playstation 2 Controller HM55B Compass
Four Servo Driver NES Controller Kalman Filter IMU
32 Servo Driver IR Remotes Memsic 2125 Accelerometer
PID Motor Control H48C 3-Axis Accelerometer Quadrature Encoders
Table 1. Propeller objects that may be of interest to robotics applications.
Figure 14. Propeller displaying accelerometer
orientation on a video display.
programming language. This makes life really easy in writing
programs for the Propeller. Instead of having to develop
code for yourself for every little task, you can use objects
that have already been developed either by yourself or
someone else to perform several functions. Table 1 shows
a short list of objects that are available for robotic
applications (downloadable from the Parallax Object
Exchange; http://obex.parallax.com). There are many
more types of objects that can be downloaded and new
ones are constantly being added. The reason I bring this up
is that it wont take you long to get some fairly advanced
robotic controls up and running.
When I started reading about all the things the Propeller
can do and looking at the source code for some of these
objects, it was pretty intimidating at first since it didnt look
like anything I was really used to seeing. So I broke down and
read the first three chapters of the Propeller manual and
typed in and ran all 12 examples in Chapter 3. This took a
couple evenings to go through. After all of this, I was quite
comfortable in reading, understanding, and writing code for
the Propeller. Remember what I said before Parallax has
some excellent documentation.
The H48C three-axis accelerometer demo setup shown in
Figure 14 is one of the coolest things I have seen in a really
long time. The ability to display information especially
animated graphics on a TV display without any special
hardware really makes visualizing where the sensor is going
intuitive. The Propeller is going to make the Nintendo WII
controller look like childs play in my opinion.
I believe the Propeller microcontroller is going to be the
big breakthrough for many of our robotic projects because
we will now have a better way to continually sense our
environment and process the various robot controls all at the
same time without any time sharing constraints that have
been holding us back. I believe this will give our robots those
lifelike responses that we have been dreaming about. SV
SERVO 03.2008 25
I hate to say this, but this is my last column as Mr.
Roboto. I have enjoyed all of your great questions and
comments over the years. It is amazing that I started this
when the magazine first came out in 2003. With our new
daughter and other time commitments at home and church,
I have decided to slow down on my writing efforts. I am not
going away completely since I will be submitting some
full-featured articles in the near future on topics that include
balancing robots, humanoid robots, and CNC controllers,
to name a few, all using and demonstrating the power of
the Propeller microcontroller. In the meantime, a highly
respected and experienced robot builder and writer,
Dennis Clark, will be taking over the Ask Mr. Roboto helm
with next months issue. Thank you for all of your support,
and keep those great questions coming in!
SIGNING OFF
IB Technologies' Micro RWD
Universal Socket Board
M
icro RWD Universal
Socket Board Universal
Base Board for ALL of the IB
Technologies Micro RWD
module types. The board
provides the antenna, power
supply regulation and RS232
/ USB interfaces to the Micro
RWD module itself.
Features include 24-pin
DIP socket for Micro RWD Modules, LEDS for visual status
indication, selectable PCB and coil based antenna for 13.56
MHz and 125 kHz operation (Antennas are etched into
the circuit board. External antennas are not necessary.),
voltage regulation allowing 9-12V DC power supplies to
be used. Additional connectors have all the Micro RWD
module electrical signals available.
The Universal Base Board data sheet details the link
settings, circuit diagram and interface connections. Gerber
files for the PCB layout (useful for the antenna dimensions)
are available on request. Please note that the Universal Base
Board does not have the USB module or a Micro RWD mod-
ule fitted as standard. These items are supplied separately.
For further information, please contact:
DC Motor Driver
T
he H-Bridge DC Motor Driver is designed to allow
TTL output to control the speed and direction of a
DC Motor. Input Voltage is 12 VDC and Rated Motor
Voltage is 4V, 6V, and 9V.
Application Note:
1) Connect Motor Pin to DC Motor. (Pin 1 & 2)
2) Connect Pin 4 to Ground and Pin 3 to 12 VDC.
3) Connect Pin 5 through 8 to any microcontroller.
Pin Assignment:
Pin1 Motor Pin (+); Pin2 Motor Pin (-); Pin3 12 VDC
Pin4 GND; Pin5 Speed Control 1; Pin6 Speed Control
2; Pin7 Motor Enable; Pin8 Motor Direction.
Racing Robot Learn
to Solder Kit
T
he Racing Robot Learn to
Solder Kit features a three
wheel motorized chassis and a
colorful robot face that flashes on and off
as the robot races out of control. Race this robot
against others to see which robot is the fastest. Kit
features 18 components that you install first to learn good
soldering techniques. Once you have mastered soldering
you then install the one IC circuitry, Face circuit, motor and
mechanical assembly. Operates from one 9V battery (not
included). Size of Face PC board about 1.5 x 2. Size of
main chassis PC board 2 x 4. Complete with all parts, PC
boards, motor and instructions. Skill Level 1.
New Products
CIRCUIT BOARDS
ROBOT KITS
MOTOR CONTROLLERS
6046461688 Fax: 6046441501
Email: fe_noble@yahoo.com
Website: www.databitz.com
Far East
Noble Sdn Bhd
Tel: 8778981005
Website: www.trossenrobotics.com
Trossen
Robotics
Tel: 8002277312
Website: www.chaneyelectronics.com
Chaney
Electronics
N
N
E
E
W
W
P
P
R
R
O
O
D
D
U
U
C
C
T
T
S
S
Is your product innovative, less expensive, more functional, or just
plain cool? If you have a new product that you would like us to
run in our New Products section, please email a short description
(300-500 words) and a photo of your product to:
newproducts@servomagazine.com
Show Us What Youve Got!
26 SERVO 03.2008
Pin1
Pin2
Pin3
Pin4
Pin5
Pin6
Pin7
Pin8
MAR08NewProd.qxd 2/6/2008 4:13 PM Page 26
The
Power SchmartModule
will power up your
circuits with your choice
of -9, -12, +2.5, +3.3,
+5, +9, and +12 volts.
$15
Very Schmart!
www.schmartboard.com
The Escape
Robot`s built-in
microprocessor
enables it to
'think on its own.
(KSR4) $29.95
20 second voice recorder/playback
module. The electret microphone is on the
board. One button records, the other button is
momentarily pressed to replay the message.
(pre-assembled) (A96010) $6.60
5mm White
water clear
LED 3.5V
10,000 mcd
(AB287)
$0.56
The Velleman Personal Scope
is not a graphical multimeter
but a complete portable
oscilloscope at the size and
cost oI a good multimeter.
(HPS10)
The robot Irog moves Iorward
when it detects sound and
repeats: start (move Iorward) -~
stop -~ leIt turn -~ stop -~ right
turn -~ stop. (KSR2) $19.95
S
o
l
d
e
r
i
n
g

r
e
q
u
i
r
e
d
Thousands
m
ore item
s on
our w
eb site!
$146.
Ask for our FREE 96 page catalog
VISIT OUR ONLINE STORE AT
www.allelectronics.com
WALL TRANSFORMERS, ALARMS,
FUSES, CABLE TIES, RELAYS, OPTO
ELECTRONICS, KNOBS, VIDEO
ACCESSORIES, SIRENS, SOLDER
ACCESSORIES, MOTORS, DIODES,
HEAT SINKS, CAPACITORS, CHOKES,
TOOLS, FASTENERS, TERMINAL
STRIPS, CRIMP CONNECTORS,
L.E.D.S., DISPLAYS, FANS, BREAD-
BOARDS, RESISTORS, SOLAR CELLS,
BUZZERS, BATTERIES, MAGNETS,
CAMERAS, DC-DC CONVERTERS,
HEADPHONES, LAMPS, PANEL
METERS, SWITCHES, SPEAKERS,
PELTIER DEVICES, and much more....
ORDER TOLL FREE
1 - 8 0 0 - 8 2 6 - 5 4 3 2
THOUSANDS OF ELECTRONIC
PA RT S A N D S U P P L I E S
Robotics Showcase Robotics Showcase
SERVO 03.2008 27
S
chmartBoard the developer of a
new technology that has significant-
ly simplified the creation of electronic
circuits for hobbyists, education, and
industry has announced the winners
of its second annual Schmartie Awards.
Schmartie Award participants, as a
part of the SchmartDeveloper program,
posted electronic circuit designs with a
bill of materials that included the correct
SchmartBoards (prototype boards) to the
companys SchmartDeveloper website.
The grand prize winner receives a $1,000
cash prize, and SchmartBoard will manu-
facture and market a SchmartModule
product with the winners name on it. In
addition, the winner will receive a com-
mission on each of these product sold.
The circuits and information about
the winners and other applicants can
be found at www.schmartdevelop
er.org. The winners of the contest are:
Grand Prize Giannis Kedros of
Thessaloniki, Greece wins the grand
prize for his Serial to USB Module.
2nd Prize Charles Wenzel of Austin,
TX wins a DSO8502 500 MHz Digital
Oscilloscope from Link Instruments for
his Low Jitter Quadrature Clock.
3rd Prize John Day of Toronto, ON
Canada wins a Weller WD1002
Soldering Station from Cooper Tools
for his USB to Serial and I
2
C Module.
Honorable Mention Daniel F. Ramirez
of Amherst, NH wins a Parallax
Boe-Bot for his Schmart DC Motor
Controller.
Honorable Mention Russell Pead of
Littleton, MA wins a Parallax
Boe-Bot for his TTL Test Board.
Honorable Mention Robert Gatt of
Port Fairy VIC, Australia wins a
Parallax Boe-Bot for his IR Proximity
Detector.
The criteria used to choose the win-
ners were originality, how well Schmart
Board technology was used in the design,
how useful the design is in the real world,
and marketability of the design.
Co-sponsors of the contest were Nuts
& Volts Magazine, SERVO Magazine,
Cooper Tools, Link Instruments, Jameco
Electronics, Mouser Electronics, Frys
Electronics, Circuit Specialists, Intellect
Lab, Parallax, and RB Technology. SV
SchmartBoard
Announces Six Winners
in Schmartie Awards
ShowcaseMAR08.qxd 2/6/2008 1:12 PM Page 27
Featured This Month:
Features
28 The Holy Grail of Combat
Robotics Usable Melty
Brain: Part 2
by Kevin Berry
31 MANUFACTURING:
Win With Bulletproof
Planetary Gearboxes
by Nick Martin
Events
34 Dec 2007/Jan 2008 Results
and Mar/Apr 2008 Upcoming
Events
35 Rumble at the Rock: BotsIQ
Gone Varsity
by Michael Bastoni
ROBOT PROFILE Top
Ranked Robot This Month:
36 Dark Pounder by Kevin Berry
28 SERVO 03.2008
I
n Part 1, we looked at four bots
that tackled the difficult task of
making the whole bot a kinetic
energy weapon using a phenome-
non of rotating bodies called
Translational Drift, or in the
combat vernacular, Melty Brain.
In a traditional Full Body Spinner
(FBS), contrary to the name, the
outer shell spins, but the base
platform is a traditional tank
drive vehicle. In a Melty, the
whole bot spins, and by varying
the drive speed of each wheel as
the bot spins, a translational
movement results.
by Kevin Berry
Usable Melty Brain Part 2:
Looking Under the Hood The
Technology of Melty Brain
THE HOLY GRAIL
OF COMBAT
ROB TICS
FIGURE 1
CombatZone.qxd 2/5/2008 5:28 PM Page 28
SERVO 03.2008 29
Ilya Polyakov of Team Carnivore
is generally credited with the first
attempts at putting this technology
in the box. He says; At the time, I
was taking Dynamics as part of my
M.E. course work and the vector
math behind combined translational
and rotational motion really hit the
spot. The identical but opposite rota-
tional velocity vectors combined with
a single translational vector made
sense in the tank-drive perfectly.
Figure 1 attempts to translate
(sorry, pun alert) this jargon into
language the mere mortal can under-
stand without having the dreaded
(and aptly named) brain meltdown.
Ilya also provided a concise sum-
mary of the systems requirements:
1) The system needs to know how
fast the robot is spinning this is
fundamental for knowing how often
to output commands.
2) The system needs to know where
the robot is pointing or where it
is within one rotation cycle, in
order to control the direction of the
translation.
3) The system must be able to make
wheel speed changes at fast rates of
20-30 Hz (once every revolution of
the robot).
4) The mechanical system must be
powerful enough to influence the
robots mass enough to make it
translate in the short time the
translational force is being applied.
Blade Runner
Polyakovs first build of Blade
Runner used a BS2SX microcontroller
and a digital compass for directional
and rotational speed sensing.
Unfortunately, the hundreds of amps
flowing through the power wires
created some very flawed compass
readings.
Version 2 was as basic as
possible with xenon strobes on the
robot illuminated once every
revolution, and Ilya acting as the
tachometer. He would sync up
the program speed in the BS2
to the robots speed manually
with a separate R/C channel,
then steer the virtual front of
the disk by advancing or
retarding the timing. This
would align the stationary
strobe light with his opponent.
Two other channels
controlled the magnitude of
the translational fwd/rev and
left/right vectors. He relates:
The system worked great in testing,
once I got the program rate
somewhat synced up to the robot
spin rate, I could controllably
translate the robot. As always,
several issues arose in the ring with
arena lighting set up for TV filming,
the strobes were barely visible, let
alone visible enough for proper
persistence of vision.
Between the difficulty of seeing
the strobe and the pressures of
combat driving, I was never able to
manually sync up the strobe. Even if I
had, I probably would not have
known it due to the lighting condi-
tions. I attribute the failure of that
version to poor planning and lack of
preparation. Had I tested the strobes
in bright sunlight, I would have
foreseen the lighting issue and pure
driving practice could have taken
care of the combat driving issues.
(Point to remember: LTFD Learn
To Flippin Drive!)
For Blade Runner 2, the Melty
system was upgraded to run on a
Rabbit microcontroller and multiple
solutions for the poor visibility were
tried. Among the more wacky ones
were strobe goggles made out of 3D
shutter goggle LCD elements, a
BS2SX, and a Linx RF serial link to the
bot, none of which were successful.
As the BattleBots TV series
wound down, so did Ilyas attempts
at creating a translational drift bot.
His path finding efforts led to other,
more successful bots.
Melty B
Rich Olson of Team SpamButcher
built a successful antweight bot
using melty brain technology. To
determine its relative position in
each spin, Rich says, Melty B uses
an accelerometer to measure the
centrifugal force created by the
rotation of the bot. The level of G
force detected is then run through a
formula that accurately determines
how fast the robot is spinning. Once
the exact spin rate is known, its
possible to determine where it is in
the current spin based on timing.
The robot flashes an LED each
time it hits a point in the spin it thinks
is forward. When spinning, the LED
appears as a streak indicating to the
driver which direction the robot will
move when the stick on the remote
is pushed forward. To adjust which
direction the LED is facing, the driver
moves the remote left or right. If a
tracking error is causing the robots
heading to drift, the driver can
compensate by steering the opposite
direction just as they would for a
car thats out of alignment.
After working the bugs out,
using just an accelerometer for
heading tracking works surprisingly
FIGURE 2. Blade
Runner 1 guts.
FIGURE 3.
Melty B guts.
30 SERVO 03.2008
well in both testing and combat. I had
originally planned on using an
infrared beacon to improve tracking
accuracy, but that doesnt seem need-
ed now. To move forward, Melty B
powers down the motor thats facing
towards the direction it wants to
travel for a portion of the spin. This, in
effect, repeatedly moves the center of
rotation slightly each spin causing
the robot to travel across the arena.
The robot controls its motors
using inexpensive Darlington drivers
(transistors). Since these only allow
for on/off control, Melty B isnt
capable of normal driving, and can
only move in a melty-brain style. The
bot has a top speed of 1,400 RPM
and can translate at about 1.5 feet
per second. At top speed, forces inside
the robot can reach over 100 Gs.
Rich coded his software over
500 lines worth in Bascom AVR.
The source code is available on his
website at www.spambutcher.
com, and is extremely well comment-
ed. He also has video of the bot in
test and combat.
The microcontroller is an
ATMEGA168 with a 20 MHz crystal.
He used the Bascoms commercial
compiler (due to program size).
Motor control is via STMicroelectronics
BU941ZT Darlington drivers from
Mouser.com, the accelerometer is a
Freescale 200G MMA2301EG, also
from Mouser. His MCU board is a
Pololu Baby Orangutan Mega168.
Scary-Go-Round
Dale Heatherington, of Dales
Homemade Robots, built and fought
a 30 lb melty at Dragoncon in 2007.
His scheme used modulated IR LEDs
both as a beacon for the bot to use
as a directional reference and to
carry control information. He used an
Atmel ATMEGA8535 microcontroller
for the brains and a 38 kHz IR
remote control receiver module for
the directional reference detector.
To monitor the directional
indicator, he built a 13 inch high
sensor tower at the center of
rotation, which has two functions.
The tower had two requirements.
First, it needed to receive infrared
control data from the handheld con-
troller transmitter (same technology
as TV remote controls), with a 360
field of view and a range of at least
20 feet. To meet this, the IR receiver is
pointed straight up into a 1/2 inch
threaded hole in a one inch OD Lexan
rod. IR light from any direction is
scattered by the threads (they act like
prisms) down towards the receiver,
giving it 360-degree coverage.
Second, it must provide a
directional reference pulse each time
the sensor under the side facing lens
is pointing at the control operator.
Without this reference pulse on
each revolution, controlled
directional movement is not
possible.
Of the two functions, Dale
found the second to be the
most challenging. IR light from
the control transmitter
reflected off nearby objects
(opponent, stage floor, and
walls) and appeared to be
coming from more than one
direction. Fortunately, light
coming directly from the
transmitter appears as a small point
source while reflections are weaker
and spread over a much larger area.
He used two sensors spaced
about 0.3 inches apart. Light directly
from the transmitter (point source)
can only focus on one at a time.
Diffused reflected light will strike
both sensors, not just one. A short
section of program code in the
microcontroller senses the condition
where one sensor is illuminated and
the other dark, times the length of
this condition, and sets a sync flag
when all conditions are true. This
pretty much eliminated falsing due to
reflections off floors and walls. He
also held off reading the sensor
again for about 270 degrees for even
more robustness.
The directional beacon receivers
homemade Lexan lens was cut from
a one inch diameter rod and then
polished. Its a cylinder lens and
focuses light only in one dimension,
converting a point image into a line.
Focal length is about 0.4 inches. The
result is very sharp focus in the
horizontal plane and no focus in the
vertical plane, so the vertical location
of the IR transmitter can vary
without affecting the performance.
As the sensor tower rotates, a
sharp vertical line of IR light sweeps
across the photo sensors. Since the
controller sends IR control signals to
the bot, it must be pointed at it at all
times. To help aim the controller, he
bolted a laser pointer on it.
Due to problems unrelated to
the melty system, Scary-Go-Round
didnt perform well. If it had, Dale
planned to add an autonomous
mode. This would consist of two
basic sensor systems, similar to those
in Sumo bots. One would sense the
edge of the stage and move away,
the other would sense the other bot
and move towards it. A complete
build report for Scary is at www.wa
4dsy.net/robot/scary-go-round.
CycloneBot
CM Robotics CycloneBot has
some of the most sophisticated
FIGURE 4. Scary-
Go-Round guts.
FIGURE 5.
CycloneBot guts.
SERVO 03.2008 31
M
ost combat bot builders in the
smaller weight classes start out
using drill motors or small custom
planetary gearbox boxes such as
those marketed by BaneBots. Many
of those builders have found, as I
have, that the stress of combat will
break the gearboxes at the most
inconvenient times.
The failure is usually a broken pin
in the second stage, which leads to
stripped gears, motor burnout, and
crushing defeat. Fortunately, this
cheap and easy upgrade will fix
the problem, win battles, and save
you money.
Parts and
Resources List
You will need only a few
consumable parts, some scrap
aluminium, and a few small tools
that you may need to buy. You will
also need some gearmotors; I chose
the BaneBots 42 mm model for this
project. For major tools, a drill press,
grinder, and either a good bench vice
or an arbor press is essential.
Jig Making
This project relies on several jigs
MANUFACTURING:
Win With Bulletproof
Planetary Gearb xes
by Nick Martin
electronics in the sport. The brain is a
Nios soft processor, running inside a
Cyclone FPGA, made by Altera. The
processor runs a Web server that is
hooked back through an 802.11
wireless link with user commands
coming from a USB joystick hooked
to a laptop. The laptop runs a
custom application which allows a
copilot to tweak various parameters
during combat.
The toughest challenge the team
hit with Cyclone Drive was maintain-
ing directional information inside the
machine. In order for the robot to
process a command to go North, it
needs to know where North is, in the
drivers reference frame. Quadrature
encoders are used on the wheels,
and this is pretty reliable for relative
direction. However, the wheels slip,
change diameter based on speed,
and when the bots go flying on
impact, all bets are off! Thus,
some system is needed for absolute
correction of heading, locking
the robots reference frame to the
drivers reference frame.
Their first approach was to use a
two axis magnetometer to measure
the Earths magnetic field. Measuring
that magnetic field in an environment
with 18 HP of brushed DC motors
that see peak currents over 1,000A
was extremely difficult and required
extensive lab testing. They had
to test a variety of different
shielding, grounding, physical
locations, wiring architectures, and
algorithm parameters for the
magnetometer to best insulate it
from the motor, motor controller, and
power line noise.
They found that on the second
version of the bot (with its lower
chassis), the magnetometer was only
2 from the arena floor, which
swamped the ability of the
magnetometer to read to Earths
magnetic field. So, they changed
over to a dual laser transmitter. One
(a visible laser) is used for sighting,
and the other (an infrared laser) for
directional information.
After prototype testing using a
simple eight bit microprocessor, they
built their beta version using a
DSP-based control system. The
fairly advanced 16-bit DSP processor
had all the usual motor control
peripherals: quadrature encoders
and pulse width modulation
generators, in particular. This setup
worked for them from the start. They
got some basic communications and
controls running, but the system was-
nt flexible enough for all their needs.
So, they moved to the combina-
tion of the Nios processor and Altera
Cyclone FPGA. The communications
system uses a dual approach:
802.11b for the primary system and
a robot-specific 900 MHz system for
the backup system.
Operation at the laptop is intu-
itive. CycloneBot can be commanded
with simple joystick input to move
north, south, east, and west.
Telemetry is passed back to the
laptop through the data stream for
real-time status checking and later
diagnosis and evaluation.
The heavy computational lifting
is handled by the onboard
Cyclone/Nios combo, freeing the
driver from the usual demanding
relative-path-based control require-
ments. This intent-based control
allows the pilot to focus his attention
completely on the strategy of the
match. SV
Material and photos were contributed by
Ilya Polyakov, Team Carnivore; Rich Olson,
Team Spambutcher; Michael Worry,
CM Robotics; and Dale Heatherington,
Dales Homemade Robots.
32 SERVO 03.2008
to provide accuracy. Start with the
output plate holder; it is nothing
more than a 3 x 3 piece of 1/2 to
3/4 MDF and a 1/4 countersink
screw as shown in Figure 1. Drill a
hole 3/4 in from each corner and
countersink the holes to fully recess
the screw head. Make sure that both
sides of the jig are completely
smooth so that the output plate
will lie flat.
The next jig holds the planetary
gears for drilling and is a little more
complicated (see Figure 2). One
corner has a 4 mm hole to help you
align the pins, while another corner
has the gear holder. When you drill
the 10 mm hole for the gear, sneak
up to the correct depth using a gear
as a depth gauge. The gear should
sit just proud of the jig surface,
which makes it easy to check that the
gear is sitting flat in the holder. I
used a slitting saw to cut the
slot, however, a hacksaw will do the
same job.
Disassemble Gearbox
Remove the long case screws
from the front of the motor and
separate the parts. You only need
the output plate for this project,
so put the remaining parts safely to
one side.
Before drilling and reaming the
output plate, remove the original
pins. You can do this with a press,
however, I just use the 3 mm screws
that hold the gearbox together as
punches and press the pins out in the
vice it is all the original screws are
really good for.
Modify the Drills
The drills double as location
devices to align the parts in the
drill press. Taper the ends of the
9/64 and 4 mm drills to about
half the original width by holding
them reversed in a cordless drill
chuck and spinning them against
FIGURE 1. The
output plate jig.
FIGURE 2. The
gear holder and
pin aligner.
Description McMaster P/N Comments
Planetary gear motor n/a
5/32 x 3/8 steel pins 98381A486 Three to four needed per gearbox
4 mm drill 28255A33 For drilling the gears
9/64 drill 8947A116 Drilling the output plate
Reamer 0.1557 2777A22 Reaming the output plate
Circlip pliers 57805A42 or similar Optional for some motors
Scrap aluminum n/a 2 x 3 x 3/8 thick for jigs
Wood or MDF n/a For drilling jigs
PARTS LIST
FIGURE 3. The major components
of the gearmotor.
SERVO 03.2008 33
the side of a fine grinding
wheel.
Modify the Output
Plate
Attach the output plate
to the MDF jig and screw it
down hard.
Insert the 9/64 drill in
your drill press with the
tapered end down. Position
the output plate jig and lower the
drill bit so the drill lines up with one
of the plate holes, then clamp the jig
in place. Reverse the drill bit and drill
out the hole, then change to the
reamer and ream out the hole to its
finished size. Repeat this sequence
for the remaining holes. Once
you have drilled an output plate,
use a fresh corner of the jig for the
next plate.
Insert a pin into the hole of the
holder jig; using an arbor press or a
sturdy bench vice, align the pin with
a hole in the output plate and press
it all the way into the plate. I find that
sticking the output plate to the vice
with a small magnet saves growing a
third hand. Make certain the pin is
exactly square to the plate before
pressing it in. Repeat for the remain-
ing holes.
Drill Out the Gears
Insert each gear into the holder
and push it down while tightening
the clamping screw, then check that
the gear is sitting flat. Align the
gear in the drill press the same way
as the output plate and drill it out to
4 mm.
After drilling out the gears,
countersink them very lightly and
polish the bores using the 4 mm drill
as shown here.
Test that the gears spin freely on
their pins; if not, then polish the
bores again. The ends of the 4 mm
pins will be sticking out from the
gears a little; grind them down
carefully until they are flush with the
tops of the gears.
Assemble the Gearbox
Clean out all traces of swarf
from the gears; it is a prime source of
jamming. Clamp the front mounting
block vertically in a vice, then drop
the output plate back over the end of
the shaft. Place the ring gear on the
front block and start inserting the
second stage gears, spinning the
output shaft to test for sticking. If
the shaft becomes harder to turn,
take out a gear and reinsert it. Once
the second stage spins freely, apply
some light grease.
Next, insert the first stage
carrier plate; I find that positioning it
above the second stage gears and
turning it until it drops down
between them works best. Test
for smooth running again before
inserting the first stage gears one by
one. Finally, close the gearbox up
with the back plate and motor and
insert all the screws another three
handed operation.
This gearbox modification will
scale up and down to different sized
gearboxes, although drilling larger
bores in small brass gears may be
risky. The next mod for a BaneBots
style gearbox should be replacing all
the stock 3 mm screws with high
tensile cap head screws and Locktite,
which will hold the parts together
reliably and will not break under
stress. SV
Nick can be contacted via his build thread at
www.robowars.org/forum/viewtopic.php
?t=74&start=0.
FIGURE 4. Pressing
out the pins.
FIGURE 5. Pressing in the
pins with the pin holder jig.
FIGURE 6. Aligning a
gear for drilling.
FIGURE 7. Polishing the gear bores. Several
sets will give you some manly calluses!
Dec. 17, 2007
Jan. 11, 2008
R
oaming Robots presented a
corporate show on December
8th at the Williams F1 HQ and
Conference Centre Christmas party,
and an educational workshop on
December 10th at St. Albans School.
W
reck-The-Halls was presented
by Carolina Combat Robots in
Greensboro, NC on December 29,
2007. Results are as follows:
Antweights 1st: Big Buzz,
Team Kelly PA.
Beetleweights 1st: Pure Dead
Brilliant, Team Rolling Thunder.
Hobbyweight 1st: Apollyon,
Team Near Choas Robotics; 2nd: Cheep
Shot 3.0, Team Rolling Thunder.
R
oboChallenge presented their
Thinktank Christmas Special
December 28th and 29th in
Birmingham, England. Results are
as follows:
Featherweight Tag Team
Competition 1st: Beauty (FP
flipper), Rip (LP flipper); 2nd: Little
Hitter (CO
2
powered Axe), G3
(FP Flipper).
Upcoming Events for
March April 2008
R
oaming Robots will present
Easter Robot Rumble on 3/22-
23/2008 at Colchester Leisure World
in Colchester, UK. Go to www.roam
ingrobots.co.uk for more details.
C
entral Illinois Bot Brawl 2008
will be presented by Central
Illinois Robotics Club in Peoria, IL
on 3/29/2008. Go to http://circ.
mtco.com for more details.
Categories will be: RC Combat (1
lb Ants only), Autonomous Sumo (3
kg, 500 g, LEGO), Line Following, Line
Maze. Entry fee is $7 per entry if you
pre-register, $10 per entry for walk-
ins. Free admission for spectators.
B
otsIQ will hold a Regional
Competition on 3/28-29/2008 in
Pittsburg, PA. Go to www.bots
iq.org for more details.
R
oaming Robots will hold an
event on 4/6/2008 at the
Fenton Manor Sports Complex, City
Road, Fenton, Stoke on Trent, ST4
2RR, UK. Go to www.roaming
robots.co.uk for more details.
R
obots Live will hold an event
4/12-13/2008 at Olympos
Burgess Hill, The Triangle, Triangle
Way, Burgess Hill, West Sussex, RH15
8WA UK. Go to www.robots
live.co.uk for more details.
S
eattle Bot Battles 2008 will be pre-
sented by Western Allied Robotics
in Seattle, WA on 4/12/2008 at the
Seattle Centers Center House. Go
to www.westernalliedrobotics.com
for more details.
Event Time: 12:00 noon - 6:00
pm, Safety Inspection: 10:00 am to
11:30 am. One and Three pound
bots. Format: Double Elimination or
Round Robin (RFL Rules). No ICE or
open flames. Entry Fee: $25 for first
3 lb or 1 lb robot. Additional robots
are half price. Special entry fee
considerations for builders who are
under 18. Arena: 8 x 8.
R
otunda Rumble will be presented
by Synergy Robotics Entertainment
in Minneapolis, MN, on 4/25-27/2008.
Go to www.kickbot.org for more
details.
Perform in front of thousands of
screaming fans at the largest tourism
destination in the world. The Mall
of America, Synergy Robotics
Entertainment, and the CRCA are
proud to announce Rotunda Rumble.
Rotunda Rumble will feature multiple
weight classes in both the STUDENT
and PROFESSIONAL leagues. The main
event will run 150 gram, Ant (1 lb),
Beetle (3 lb), Hobbyweight (12 lb),
15lb, and Featherweight (30 lb) classes.
Prizes: Student 15 lb Class;
$2,500 in merchandise for 1st Place,
Trophies to 1st, 2nd, and 3rd Place.
Professional 12 lb and 30 lb Class;
$500 to 1st Place, Trophies to 1st,
2nd, and 3rd Place.
B
otsIQ: The Competition 2008 will
be presented by BotsIQ in Miami
Beach, FL on 4/30/2008-5/4/2008. Go
to www.botsiq.org for more details.
All teams are welcome to
compete in the following categories:
Table Top (Task oriented same
game as last year); 15 pound
competition Middle Schools, High
Schools, and Post Secondary
Institutions; 120 Pound Competition
High School and Post Secondary
Institutions. SV
34 SERVO 03.2008
EVENTS
Results and Upcoming Events
T
he Plymouth North High and
Plymouth South High School
Robotics Teams sponsored a BOTSIQ
15 lb competition. The Plymouth
North High school engineering class,
under the direction of teacher
Michael Bastoni and welder James
Stevens, designed and constructed a
12 Octagonal battlebox. The 1/8
steel floor is supported by eight sec-
tions of welded 3 structural channel
aluminum frames. The walls are 5x5
x 3/8 polycarbonate framed by
1-1/4 aluminum square tube frames.
The roof is 3/8 poly and aluminum
set on an approximately 1:6 slope.
This is the first BOTSIQ arena of this
design in the country and it was
shipped to Florida and used at the
2007 National BOTSIQ championship.
The Event
Following the 2007 BOTSIQ
championship, we announced the
desire to host a combat robot event
in Plymouth, MA. We were not sure
how many robots would come since
we made the announcement in the
early weeks of October 2007, listing
December 1, 2007 as the event date.
We had hoped for 6-10 robots and
expected that we could grow the
event from there. We were amazed
that 25 15 lb combat robots from 18
schools showed up ready to fight!
We had only one forfeit throughout
the whole day of combat.
The event went off without a
hitch. The arena
was set up in the
center of the manu-
facturing shop sur-
rounded by lathes,
mills, and welding
equipment, a per-
fect place for a
combat robot event
and an homage to the early
days of Marc Thorpe and the
West Coast crazies who
birthed the sport. We set up
a 14 TV screen in the
auditorium and pumped in
live video and sound from
inside the arena so folks
could sit in movie theater
comfort and hear and watch
15 lb fighting robot action,
which looked and sounded
like 300 lb combat robots on the big
screen. The live and video visual
experience was enhanced since we
shut down the room lights and lit
only the BattleBox, and pumped
amplified sound into the arena room,
as well as the auditorium. Little bots
got real big!
We had both high school and
college teams from as far away as
Florida and New Jersey.
Results
First Place (Undefeated): Good
Knight from Bergen County
Technical School, NJ.
Second Place: The Hook from
WI.
Third Place: Juggernaut from
Worcester Polytechnic Institute, MA.
I think the story here is the grass
roots start up. Students designing
and building the arena AND the
robots, and making the venue
available to surrounding teams. In
this way, engineering in general, and
robotics in particular, can become in
time varsity level competitive sports
that will rival the status of other var-
sity sports programs. We didnt nail a
peach basket to an auditorium
balcony ... we built a combat robot
arena! This event will be held on a
semiannual basis, made possible
through the sponsorship of Entergy
Corporation (www.entergy.com),
SolidWorks Corporation (www.solid
works.com), and GEARS Educational
Systems (www.gearseds.com).
The next Rumble at the Rock is
scheduled for March 8th, 2008. SV
SERVO 03.2008 35
RUMBLE AT THE R CK:
BotsIQ Gone Varsity
by Michael Bastoni
Lara Spinelli,
driver for bot
Good Knight
of the Bergen
County
Academy
robotics team.
The 12 Octagonal arena was designed and
built by the organizers, sponsors and
students at Plymouth North High School
Engineering Program. Copies of these plans
made by Mahuta Tool Corp (www.mahuta
tool.com) are available from BOTSIQ by
contacting Nola Garcia at nola@botsiq.org
36 SERVO 03.2008
D
ark Pounder has competed in
RobOlymics/RoboGames 2007,
Smackdown in Sactown III, Texas
Cup, Robot Shoot-Out, Most Extreme
Robot Challenge, Robot Rebellion
5.4, Robot Rebellion 5.3, Robot
Rebellion 5.2, Robot Rebellion 5.1,
2004 RFL Nationals, Rocket City
Robot Assault, Robot Rebellion 7.0,
Robot Rebellion 6.0, and Robot
Rebellion 4.0. Details are listed
below:
Frame: 3/16 7075 aluminum
vertical supports, 0.03 titanium belly;
horizontal carbon rod stiffener, 0.03
titanium formed shell.
Drive: Two BaneBots 11:1 ratio 16
mm drive motors to 3/4 wheels.
Wheels: Two 1 Dubro wheels
turned down to 3/4.
Configuration: Vertical asymmetric
bar spinner with rounded form wedge
(front or rear, depending on opponent).
Drive ESC: Barrello Ant 150 dual
5A controller.
Drive batteries: Two 2S1P 400
mAh 20C Hyperion LiPoly.
Weapon: 3 asymmetric cleaver
blade of 1/8 hardened chromoly steel;
drive pulley mounted to blade on live
axle (3/8 grade 8 bolt), 27,000 RPM.
Weapon power: 16A start-up and
8A cruise at 14.8V.
Weapon motor: GWS inrunner
brushless motor GWBLM005A
(3,900 Kv).
Weapon ESC: GWS 25A brushless.
Armor: 0.03 6Al-4V titanuim
shell form wedge, 0.06 Ti vertical
armor near the blade.
Future: Design works, but will tune as
necessary, 1.5 ounces under weight, so
some future armor additions possible.
Design philosophy: Rounded form
shell maximizes armor strength,
protected wheels, blade rotation, and
bot orientation direction dependent
on exploitation of opponents weak-
nesses; assymetric blade for greater
bite on opponent. Reliable, pre-
dictable vertical spinner weapon, low
center of gravity with high velocity
weapon of smaller diameter. SV
Photos and information are courtesy of
Russ Barrow. All fight statistics are courtesy of
BotRank (www.botrank.com) as of January 12,
2008. Event attendance data is courtesy of The
Builders Database (www.buildersdb.com) as of
January 12, 2008.
ROBOT PR FILE
by Kevin Berry
TOP RANKED ROBOT THIS MONTH
Weight
Class
Bot Win/Loss
Weight
Class
Bot Win/Loss
150g VD 26/7 150g Micro Drive 7/1
1 lb
Dark
Pounder
44/5
1 lb
Dark
Pounder
28/3
1 kg Roadbug 24/10 1 kg Roadbug 14/5
3 lb 3pd 48/21 3 lb Limblifter 15/3
6 lb G.I.R. 14/2 6 lb G.I.R. 11/2
12 lb Solaris 42/12 12 lb Rants Pants 11/2
15 lb Humdinger 26/4 15 lb Humdinger 26/4
30 lb Helios 31/6 30 lb Billy Bob 8/0
30 lb Bounty
Hunter
9/1
30 lb
(sport)
Bounty Hunter 9/1
60 lb
Wedge of
Doom
43/5 60 lb Agent 7 5/0
120 lb
Devil's
Plunger
53/15 120 lb Touro 5/0
220 lb Sewer Snake 35/9 220 lb Brutality 4/0
340 lb Shovelhead 39/15 340 lb
Psychotic
Reaction
4/1
390 lb MidEvil 28/9 390 lb MidEvil 3/0
Top Ranked Combat Bots
Rankings as of January 14, 2008
Historical Ranking is calculated by
perfomance at all events known to
BotRank
Current Ranking is calculated by
performance at all known events,
using data from the last 18 months
History Score Ranking
(sport)
Historical Ranking: #1
Weight Class: 1 lb Antweight
Team: Dark Forces
Builder: Russ Barrow
Location: Richardson, TX
BotRank Data Total Fights Wins Losses
Lifetime History 49 44 5
Current Record 31 28 3
Events 14
Dark Pounder Currently Ranked #1
http://RoboGames.Net
June 12-15th, 2008
San Francisco, CA
Events:
Compete at RoboGames 2008!
Last year, over 1000 builders from around the world brought over 800
robots to San Francisco, in the 4th annual international event. This year,
we expect even more robots and engineers to compete. Be one! With 80
different events, theres a competition for everyone - combat, androids,
sumo, soccer, Lego, art, micromouse, BEAM, or Tetsujin! More than half
the events are autonomous. Even if you just come to watch, youll be
overwhelmed with the diversity.
Last year, RoboGames hosted teams with over 800 robots from Argentina,
Australia, Austria, Brazil, Canada, China, Colombia, Czech Republic,
Denmark, Germany, India, Indonesia, Iran, Japan, Korea, Mexico, Nether-
lands, Peru, Singapore, Slovenia, Sweden, Taiwan, UK, and the USA.
Be a RoboGames Sponsor!
RoboGames is the worlds largest open robot competition - letting people
of any age, gender, nationality, or affiliation compete. Sponsoring Robo-
Games not only helps more people to compete, but also gets your
company unrivaled press coverage and visibility. The event has been
covered by CNN, ESPN, Fox, CBS, ABC, NBC (live), EBS Korea, NHK
Japan, BBC, and countless print and web companies. Your logo can be
everywhere the cameras turn!
Rent a Booth!
Booth spaces are at the front of the venue, ensuring lots of traffic. With
3000-5000 people each day, youre company will get amazing traffic!
-SERVO Magazine
Full Page.qxd 2/6/2008 2:48 PM Page 37
38 SERVO 03.2008
I
t is important that a power-assist exoskeleton robot
automatically assists the users motion according to
that motion intention in real time. Electromyographic
(EMG) signals which are generated when muscles are
activated are one of the most important biological
signals to determine the users motion. The amount of
the EMG signal indicates the muscle activity level (i.e.,
the amount of generating force) and it can be easily
measured using simple electrodes.
If the amount of generating force by certain
muscles is estimated, the amount of users joint torque
can also be estimated (see Figure 1). Therefore, it can be
used to activate the power-assist exoskeleton robot
automatically, since it directly reflects the intention of
the user. Consequently, human motion can be estimated
if the amount of muscle force of certain muscles is
estimated. If the users motion is estimated in real-time,
the motion can be easily assisted by the exoskeleton.
The EMG-based control is not very easy to be realized,
however, because of several reasons.
In this article, EMG-based control methods for
power-assist exoskeleton robots will be introduced. Soft
computing technologies such as fuzzy reasoning, neural
networks, or genetic algorithms are powerful tools to
make the robot system intelligent. They can be applied
to develop an effective EMG-based controller for power-
assist exoskeleton robots. We will discuss, two kinds of
EMG-based control methods in which soft computing
technologies are introduced and applied.
The Geometry of Power-Assist
Because the power-assist exoskeleton robot is
supposed to be directly attached to the users body, the
design condition of the robot architecture is restricted in
comparison with that of ordinal robots. The actuators,
sensors, links, and frames of the exoskeleton must all be
located outside of the users body and not disturb the
users motion under any configuration. Moreover, the
weight of the exoskeleton should not be directly
supported by the users body.
Therefore, the hardware design is harder to come
up with than that of ordinary robots. Development of
small-size, light-weight, and high-power actuators such
Control of Power-Assist
With Biological Signals
A viable power-assist exoskeleton
robot sometimes called a power
suit, man amplifier, or man
magnifier is something that
many people in industry, military,
and medicine, have been anxiously
waiting for. Recent progress in
robotics technology has increased
the number of power-assist
exoskeleton robots that have come
onto the scene.
Biceps
(agonist muscle)
Triceps
(antagonist muscle)
Generating muscle force
for elbow flexion
Flexion
EXOSKELETON
ROBOTS
by Kazuo Kiguchi
FIGURE 1. Joint torque
and muscle force.
Kiguchi.qxd 2/5/2008 6:05 PM Page 38
as artificial muscles are required to be
practically wearable for daily living.
Upper-limb motion is involved in
many important activities in daily living,
so assistance here is important for
physically weak persons. Figure 2
shows an example of 4 DOF (degrees
of freedom) upper-limb motion (i.e.,
shoulder vertical flexion/extension,
shoulder horizontal flexion/extension,
elbow flexion/extension, and forearm
pronation/supination) [1].
In this case, the exoskeleton robot is
attached to the mobile wheelchair.
Therefore, the user does not carry any of
the weight at all. The exoskeleton main-
ly consists of a shoulder motion support
part, an elbow motion support part, and
a forearm motion support part.
The shoulder motion support part
is composed of an upper arm link, driv-
er and driven pulleys (one for shoulder
horizontal flexion/extension motion,
another one for shoulder vertical flex-
ion/extension motion), two DC motors,
two potentiometers, an arm holder, and
the mechanism for the center of
rotation (CR) of the shoulder joint. The
1 DOF elbow motion part consists of a
forearm link, pulleys, a DC motor, and a
potentiometer. The forearm motion sup-
port consists of a wrist frame, an inner
and outer wrist holder, a wrist cover, a
wrist force sensor, and potentiometers.
Usually, the movable range of the
human shoulder is 180 in flexion, 60 in
extension, 180 in abduction, 75 in
adduction, 100-110 in internal rotation,
and 80-90 in external rotation. The lim-
itation of the movable range of the fore-
arm motion is 50-80 in pronation and
80-90 in supination, and elbow motion
is 145 in flexion and -5 in extension.
Considering the minimal amount of
motion required in everyday life and the
safety of the user, shoulder motion of the
4 DOF exoskeleton is limited to 0 in
extension and adduction, 90 in flexion,
and 90 in abduction. Limitation of the
forearm motion is 50 in pronation and
80 in supination, and 120 in flexion and
0 in extension for the elbow motion.
In order to activate the exoskele-
ton in accordance with the users
intended motion, the EMG-based con-
trol can be applied as explained next.
In order to control the 4 DOF upper-
limb motion, 12 kinds of EMG signals
should be used, as shown in Figure 3.
Control with Biological
Signals
In order to assist the motion of the
user, the exoskeleton robot must deter-
mine the generating motion in realtime.
The users motion can be estimated in
real time by monitoring the EMG signals
of the certain muscles. Since the
amount of EMG signal indicates the
activity level of the muscles, the amount
of generating force by the user can be
estimated by monitoring these signals.
However, this EMG-based con-
troller is not very easy to be realized,
because: 1) obtaining the same EMG
signal for the same motion is difficult
even with the same person since the
signal is biologically generated; 2) the
activity level of each muscle and the
way theyre used for certain motion is
different among individuals; 3) activity
Cont r ol of Power - Assi st Exoskel et on Robot s wi t h Bi ol ogi cal Si gnal s
SERVO 03.2008 39
FIGURE 2. 4DOF power-
assist exoskeleton.
FIGURE 3. Location of electrodes.
40 SERVO 03.2008
of antagonist muscles affects the joint
torque (see Figure 1); 4) many muscles
are involved in a joint motion; 5) a
muscle is used for more than one kind
of motion; 6) the role of each muscle
for a certain motion varies with joint
angles; and 7) the activity level of some
muscles (such as bi-articular) are affect-
ed by the movement of the other joints.
There are basically two kinds of
methods to carry out power-assistance
based on the users EMG signals. One
way is a fuzzy-neuro control method
(the combination of flexible fuzzy
control and an adaptive neural
network) [1]-[4] and the other is a
muscle-model based control method.
In the first method, the users motion
is estimated based on the EMG activation
patterns of the related muscles, and then
the power-assist is performed to accom-
plish the estimated motion. The fuzzy
IF-THEN rules are designed based on the
relationship between the human motion
and the EMG activation patterns of the
related muscles. The designed fuzzy
IF-THEN rules are transferred into the
form of neural networks so they can
adapt to an arbitrary user. As the
number of assisting DOF is increased, the
required fuzzy IF-THEN rules become
more complicated to cope with this.
In the second EMG-based method,
the users motion is estimated on the
amount of EMG activity levels of the
related muscles. A muscle-model (i.e., a
matrix) that relates human joint torque
and the amount of EMG activity levels is
based on the knowledge of human anato-
my. However, each component of the
matrix must be modified according to the
posture of the user, since the relationship
between the human joint torque and the
amount of EMG activity levels varies.
A fuzzy-neural network can be
applied to modify the muscle-model in
real time according to the posture of
the arbitrary user. The control system
for the power-assist exoskeleton robot
for 5 DOF upper-limb motion
(i.e., shoulder flexion/extension, shoul-
der adduction/abduction, shoulder
internal/external rotation, elbow
flexion/extension, and forearm prona-
tion/supination) is shown in Figure 4.
The relationship between the EMG
signals and the generated joint torques
are written as the following equation if
the posture of the users upper-limb
does not change.
EQUATION 1:
where
sv
is torque for shoulder flex-
ion/extension motion,
sh
is torque for
shoulder adduction/abduction motion,
sr
is torque for shoulder rotational
FIGURE 4. Architecture
of proposed EMG-based
controller.
16
15
2
1
16 15 2 1
16 15 2 1
16 15 2 1
16 15 2 1
16 15 2 1
Ch
Ch
Ch
Ch
w w w w
w w w w
w w w w
w w w w
w w w w
f f f f
e e e e
sr sr sr sr
sh sh sh sh
sv sv sv sv
f
e
sr
sh
sv

Abduction Moving a limb away from
the midline. Think B for bird (aBduction)
raising your arms like a bird preparing
for flight.
Adduction Moving a limb toward
midline. Think D for down (aDduction)
pushing your arms down in a resting
position.
Pronation Rotating the forearm and
hand so that the palm is down. Think P
for pouring water out of an imaginary
bowl in the palm of your hand.
Supination Rotating the forearm and
hand so that the palm is up. Think S for
holding a bowl of soup in the palm of
your hand.
Anatomy Lesson
motion,
e
denotes torque for elbow
motion, and
f
is torque for forearm
motion. Chn is the EMG signal meas-
ured in channel n, w
svn
is the weight for
nth EMG to estimate the torque for
shoulder flexion/extension motion,
w
shn
is the weight for nth EMG to
estimate the torque for shoulder
adduction/abduction motion, w
srn
is
the weight for nth EMG to estimate the
torque for shoulder rotational motion,
w
en
is the weight for nth EMG to esti-
mate the torque for elbow motion, and
w
fn
is the weight for nth EMG to esti-
mate the torque for forearm motion.
Five joint angles (shoulder flex-
ion/extension angle, shoulder adduc-
tion/abduction angle, shoulder inter-
nal/external rotational angle, elbow
flexion/extension angle, and forearm
pronation/supination angle) are used
as inputs for the fuzzy-neural network.
Each joint angle is divided into three
regions (i.e., FL: flexed region; IM:
intermediate region; and EX: extended
region for shoulder flexion/extension,
adduction/abduction angles, and
elbow flexion/extension angle; and IN:
internal region; CE: center region;
and EX: external region for shoulder
internal/external rotational angle and
forearm pronation/supination angle).
The output from the fuzzy-neural
network is the coefficient for each
weight in Equation 1. Consequently,
the muscle-model (the weight matrix) in
Equation 1 is adjusted by the fuzzy-
neural network at every sampling time
during the control as shown in Figure 4.
The fuzzy-neural network is
trained to minimize the difference
between the estimated torque and
the measured generated torque during
the training period using the error
back-propagation learning algorithm
(i.e., one of the most typical learning
algorithms for neural networks).
Evaluation
In order to evaluate the effective-
ness of the motion estimation methods
for the power-assistance explained here,
SERVO 03.2008 41
FIGURE 5. Location of
markers in the experiment.
FIGURE 6. Generated joint
torque (Subject A).
0 10 20
- 10
0
10
T
o
r
q
u
e

(
N
m
)
Time (sec)
0 10 20
- 10
0
10
T
o
r
q
u
e

(
N
m
)
Time (sec)
0 10 20
- 5
0
5
T
o
r
q
u
e

(
N
m
)
Time (sec)
0 10 20
- 5
0
5
T
o
r
q
u
e

(
N
m
)
Time (sec)
0 10 20
- 0.01
- 0.005
0
0.005
T
o
r
q
u
e

(
N
m
)
Time (sec)
42 SERVO 03.2008
experiments have been carried out to
determine the relationship between the
human 5 DOF upper-limb motion
(3 DOF shoulder joint motion, elbow
joint motion, and forearm pronation/
supination motion) and the EMG signals
of the related muscles (deltoid-anterior
part, deltoid-middle part, deltoid-
posterior part, pectoralis major-c
lavicular part, pectoralis major-lateral
part, biceps-medial part, biceps-lateral
part, triceps-medial part, triceps-lateral
part, teres major, teres minor, infraspina-
tus, pronator teres, flexor carpi radialis,
anconeus, and supinator). The effective-
ness of the fuzzy-neuro method has
been proven in many papers [1]-[4].
In one experiment, random upper-
limb motion is performed by five
healthy, elderly male persons (all of
them are 65 years or older), and the
estimation error of the generated
upper-limb motion is evaluated. The
motion of each subject is measured by a
3D motion capture system (Hawk Digital
System and a high-speed camera from
NAC Image Technology, Inc.). In order to
capture the 3D motion, 17 markers are
put on each subject in the experiment as
shown in Figure 5. The motion of the
subject is monitored with eight cameras.
Each joint torque generated by
each subject during the experiment is
estimated from the motion results
calculated using human models [5].
The experimental results of the
torque generated by subject A are
shown in Figure 6 as an example. Figure
7 shows the error between the generat-
ed torque by subject A and the estimat-
ed torque with and without the second
EMG-based control method. It can be
seen that the amount of the torque error
becomes almost zero when the weight
matrix is adjusted by the second EMG-
based control method. Similar results
were obtained with the other subjects.
These results show that the users
joint torque can be estimated and a
certain percentage of the estimated
torque would be assisted by the
exoskeleton robot in real time. Thus,
the power-assist is realized with the
proposed EMG-based control method.
Ideas for Future Expansion
Recent progress of robotics tech-
nology has brought several exoskeleton
robots onto the market. However, they
are still not ready for daily consumer
use. More advanced soft computing
technologies need to be developed to
make the power-assist robots more
intelligent and flexible. SV
[1] K. Kiguchi, Active Exoskeletons for
Upper-Limb Motion Assist, Journal of
Humanoid Robotics, vol. 4, no. 3, 2007.
[2] K. Kiguchi, S. Kariya, K. Watanabe, K.
Izumi, T. Fukuda, An Exoskeletal Robot
for Human Elbow Motion Support
Sensor Fusion, Adaptation, and Control,
IEEE Trans. on Systems, Man, and
Cybernetics, Part B, vol.31, no.3, pp.
353-361, 2001.
[3] K. Kiguchi, T. Tanaka, T. Fukuda,
Neuro-Fuzzy Control of a Robotic
Exoskeleton with EMG Signals, IEEE
Trans. on Fuzzy Systems, vol.12, no.4,
pp.481-490, 2004.
[4] K. Kiguchi, R. Esaki, T. Fukuda,
Development of a Wearable Exoskeleton
for Daily Forearm Motion Assist,
Advanced Robotics, vol.19, no.7, pp.
751-771, 2005.
[5] H. Okada, M. Ae, N. Fujii, Y. Morioka,
Body Segment Inertia Properties of
Japanese Elderly, Biomechanism 13,
pp.125-139, 1996 (in Japanese).
References
0 10 20
-10
0
10
Time (sec)
T
o
r
q
u
e

(
N
m
)
Before learning
After learning
0 10 20
-0.1
0
0.1
Time (sec)
T
o
r
q
u
e

(
N
m
)
Before learning
After learning
0 10 20
-10
0
10
Time (sec)
T
o
r
q
u
e

(
N
m
)
Before learning
After learning
0 10 20
-10
0
10
Time (sec)
T
o
r
q
u
e

(
N
m
)
Before learning
After learning
0 10 20
-10
0
10
Time (sec)
T
o
r
q
u
e

(
N
m
)
Before learning
After learning
FIGURE 7. Estimated joint torque
error (subject A).
O
ne goal of this project was to
re-use as much existing stuff as
possible and to give leftover parts a pur-
pose. I felt it appropriate to use recycled
components, too. When building some
one off projects and hand wired PCBs,
I like to try and re-use old parts that may
have otherwise ended up in a landfill.
Some of the parts youll see this
time came from some unlikely places.
In particular, some of the relays used
on the board to multiplex the sonar
sensors came from old Token Ring
network cards that were being thrown
out. The small LEDs used on that same
board came from an old ADIC tape
library controller board that also was
heading out to the dumpster.
This is a great use of parts when pro-
totyping. For projects that will need more
than a couple of like boards, I would go
ahead and layout a new PCB instead of
hand wiring them all on a proto board.
Id also make sure that any parts used are
ones that are readily available.
How is all this stuff going
to end up in BOB?
As stated in the first
article, anything that is going
in the robot is simply going to
bolt-in or plug into the exist-
ing chassis and electronics.
No alterations to the original
robot are going to be done.
This is one of the first issues
to overcome. Where and how
am I going to mount all this
new stuff in BOB? To start,
there is the main Handy
Board controller, the H-bridge
driver for the main drive motors, the
custom power board, the Polaroid Sonar
ranger, and a Sonar MUX board. There
will also be a few other boards which will
be covered in the next article.
The best place to mount everything
Last month, you read about how I started re-using
some leftover robotics components to revive an old
Androbot BOB robot chassis. I focused on what was
needed to control the drive motors in the BOB base. I
also covered upgrades to an old Handy Board controller
to make it more suitable for use on a larger robot and
enhancements to the board to make it more reliable. Since
the only thing I had to start with was the empty robot shell
and drive motors, this entire project is covering the rest of the
parts that are needed to build a functional autonomous robot. If it werent
for the re-use of the body shell and drive assembly, this series would basically cover building a
whole new robot from scratch. During the process, Ill try to illustrate how each phase was going
to work out, how plans change along the way, and the resulting changes that were done to
accommodate those changes to make it all work.
How
everything
mounts to
the base.
P A R T 2
Reviving an
Androbot BOB
SERVO 03.2008 43
by Robert Doerr
Doerr2.qxd 2/6/2008 1:46 PM Page 43
44 SERVO 03.2008
seemed to be right on the top of the chas-
sis like the original TOPO controller board
mounted to its base in the RF model (see
the November 07 SERVO). I believe that
most of the original boards would have
mounted in this area. I didnt want to drill
into the chassis to mount the boards.
Instead, I wanted to have a separate plate
which I could secure all the electronics to
and that plate could bolt to the existing
mounting studs on the chassis.
After taking a look at the base, I
headed out to the garage to sort
through some old aluminum plates I
had been saving. I found a rectangular
aluminum panel (salvaged from some
odd piece of computer gear) which
looked like it would do the job.
Unbelievably, it was just the right size!
It already had a hole drilled in each cor-
ner and those almost lined up perfectly
with the mounting studs on the base. I
only had to drill out the holes a little
more to use them and bolt it in place.
That plate was pulled back out and
placed on the workbench to figure out
how to secure all the electronics to it.
First, I just set all the boards out on the
plate to determine what locations made
the most sense for the placement of
each board. The final result had the
dual LMD18200 based H-bridge board
near the middle with the Handy Board
next to it toward the back. The portion
in front of the H-bridge is reserved for
a co-processor and speech, another
H-bridge for the head, and sound amp
which will be discussed next month.
Behind the Handy Board is the
power supply and distribution section.
This includes a 24V to 5V DC/DC convert-
er, barrier strip, and custom power board
to handle the 6V and 9V power needed
by some sections of the robot. There
really wasnt a place for the Polaroid
sonar module or the custom board to
multiplex the five sonar transducers.
For that, I used a set of tall stand-
offs and another salvaged aluminum
part. This was a cover that was recycled
from an old video to fiber adapter. It
seems to shield the sonar and keep it
from interfering with the Handy Board
and the other electronics.
Power is provided from a pair of
12V 7 Ah batteries just like the TOPO
robots. These are going to power both
the logic and the drive system. The robot
drive motors are meant to run at 24V,
but the logic and everything else runs at
lower voltages. A choice had to be made
right here. Was everything going to run
off 24V and be stepped down to a
useful level, or would it be better to tap
off 12V from the first battery? Other
voltages can be derived from these.
The disadvantage of this method is
that the two batteries may discharge at
different rates since the first battery
will have more load on it. If the charg-
ing system can handle it and charge
each battery separately, then this can
be an easier way to go. Since the
power for the overall robot would be
switched from 24V, I was going to use
a large relay that would switch on the
12V supply from the center tap so that
would be off when the robot was off.
I built this section on a proto board
and it worked fine. However, before
implementing this solution I changed
my mind. I had some extra blank PCBs
for the TOPO II battery monitor board
and I built one to use on BOB. This has
several advantages. It acted correctly
for this type of robot (one switch to
push on and another to push off). It
also handled charging of the two
batteries, and would automatically shut
down the robot when the battery
power was drained. The input had both
the full 24V and the 12V from between
the two batteries and an onboard relay
would switch them both on at once.
Now, the extra relay to turn on the
12V supply on the custom power
board was no longer needed. Instead
of just yanking it off the proto board, I
just re-wired it so that it can be directly
controlled from the Handy Board or
extra co-processor board. It will be
available for future expansion to switch
on and off a large 12V or 24V load,
depending which is eventually needed.
Whether designing a full blown
commercial project or just wiring up a
homebrew one from leftover parts, it is
important to think about safety. A lot of
people ignore this and may get away
with it for a while. It is best to engineer
in at least some sort of protection for
your circuits. On the output of each one
of the batteries, a fuse is installed to
provide short circuit protection. It is much
better that a fuse should melt instead of
your wiring harness! The Handy Board
has a Polyfuse that acts as a safety in
case you overload that particular board
REVIVING AN ANDROBOT BOB: Part 2
Original TOPO battery monitor board.
The custom power board (top). The custom power board (bottom).
or any circuits taking power from the
Handy Board itself. Power for the Handy
Board, expansion board, Sonar Multiplex
board, and encoders all come from a
commercial DC/DC converter module in
a nice compact DIP-24 pin format. These
are available from several manufacturers
(CINCON, AZTEC, etc.) with a variety of
options for input and output voltage. The
advantage of using a module like this is
that they are usually very efficient (80%-
90% range), have built-in short circuit
protection, and have good regulation.
The one used for this robot
accepts 18V-36V in and puts out a
steady 5V. It made more sense to go
this route instead of spending a lot
of time designing an entire power
supply. This DC/DC converter gets
power as soon as the main 24V from
the batteries are switched on.
The custom power supply module
takes both 12V and 24V into the board.
The 12V runs through four 1N5401
diodes in series to provide approximate-
ly 9V out for the audio amp and SX48
co-processor board. It also feeds a 6V
voltage regulator for a dedicated 6V out
used specifically for the Polaroid sonar
ranger. The 24V power in will be avail-
able to check the overall system voltage
and is used to power a relay that the
Handy Board can control to switch on
and off a 12V supply to power some
device (like a vacuum) later on.
The voltages that are available
when the robot is switched on (12V, 9V,
6V, etc.) are each brought out to a bar-
rier strip. This helps keep the wiring neat
and allows for easy changes later on.
Running all those sonar
transducers
Instead of installing a Polaroid sonar
ranging board for each sonar transducer
(five boards), it made more sense to just
use a single board and then add some
additional circuitry to allow it to
drive each one of the five transducers.
A proven way to do this is to use
small relays to switch the appropriate
transducer in circuit. It is important that
there is always one transducer selected
SERVO 03.2008 45
Power wiring. Labels for power board.
Schematic of custom
power board.
46 SERVO 03.2008
and only one at a time. The schematic
presented here will do just that.
It only requires 5V, ground, and
three data lines to select the appropriate
relay. It uses a 74HC138 and a 74HC14
inverter to select, and a ULN2803 to
drive the relays. To determine which
relay will be energized, we must decode
the signals for relay select. This is what
the 74HC138 chip does well.
It is a three to eight decoder and
depending upon what pattern is applied
to its inputs A0, A1, and A2, you will
get one of the outputs (Y0 through Y7)
low and the remaining will be high. We
will use the active low to activate a relay
for a particular transducer.
That will be great for the first five
(0-4), but what happens when the bit
pattern is accidentally set to 5-7?
Remember that one transducer and only
one must always be selected. That is
why we take the signal for A2 to switch
the last relay and ensure that all others
are switched out of the circuit. The
inverters from the 74HC14 convert the
active low to become an active high.
The active high goes to the ULN2803
chip which drives the relay. It contains a
set of eight Darlington transistors and
has protection diodes built in.
As a visual indication, a small LED
was placed next to each relay so it is
easy to tell when it is energized. It only
takes wiring a small LED and current
limiting resistor across the coil of each
relay. It is best to select a resistor for
the LED so that it can be clearly seen
but not overly bright. Anything more is
Sonar MUX board (bottom). Sonar MUX board (top).
Schematic of custom
sonar MUX board.
just wasteful. In this case, a 560 ohm
resistor worked out well.
For the most part, I had used 74
series TTL chips in projects and hadnt
used CMOS chips in new projects. The
Handy Board is all CMOS based for low
power, so I kept that convention when
making the logic for the Sonar MUX
board. One thing that comes up with
CMOS is to never leave any inputs
unconnected! This is something you
must remember when working with
CMOS chips.
Everything was wired up fine for the
first test but I did it on the bench without
connecting all three selection lines to the
Handy Board. When power was applied,
all five of the relays and the associated
LEDs went crazy! This happened because
an unconnected CMOS input can oscil-
late wildly and as a result the relays
would energize at random. This wouldnt
be a problem as long as it was connect-
ed to the Handy Board or some other
controller since the input would not float.
To make it work on the bench and
also in circuit, I just added three 10K
pull-down resistors on each data selec-
tion line. This ensured that no lines
would be un-connected and would
default each line to a known state. It
took care of the problem nicely. If for
some reason one of the selection lines
goes open, it will always be low at the
sonar MUX board. Otherwise, if a lead
opens up, the relay board could start
acting up and cause problems.
When I went to connect the
Polaroid sonar ranger to the expansion
board, I got quite a surprise. The Handy
Boards expansion had a nine-pin connec-
tor. The docs that came with the module
also showed a similar nine-pin connec-
tion. Should be easy to connect, right?
(Well if it was, I wouldnt have had as
much to write about.) As it turns out, the
module I was going to use only had an
eight-pin connector. Since pin 9 is nor-
mally used, I had to find the appropriate
pinout for the eight-pin module and then
wire up the connection to the Handy
Board to compensate for the difference.
After some online research of the
Polaroid sonar modules, I found the
details I needed to wire it up. A table
showing the mapping for the nine-pin to
eight-pin modules (at least for the
module I was using) is provided in this
article. Since I didnt have that odd
nine-pin connector on the Handy Boards
expansion, I just used a length of ribbon
cable to make the connection.
There are five lines needed to drive
the ranger. I soldered one end of the
cable directly to the expansion board
for GRN, INIT, ECHO, and BINH. A fifth
line for +6V power came from a dedi-
cated voltage regulator on the custom
power board that was specifically there
for the ranger. These all went to an
eight-pin .100 inch straight connector
to plug into the module I was using.
If you do a similar project, you can
avoid all this by using one of the newer
SensComp (replacement for Polaroid)
ranging modules which already has a
matching nine-pin connector for the one
used on the Handy Board. The only
concern is with the power supply for the
sonar if you start altering the Handy Board
power supply like I did on this project. You
need to be sure it gets reliable power.
Table 1 shows pinouts for common
nine-pin and eight-pin Polaroid sonar
modules. NOTE: Before trying out any
sonar modules like this, make sure that
the transducer is connected to the
ranging board or you can damage the
ranging board.
WARNING: When the Polaroid
sonar ranging board is in operation,
the voltage applied to the transducer
can reach several hundred volts! Even
though Ive been aware of this and
worked around these types of sensors
for years, I finally got nailed when
testing this. It only stung for a minute,
but that will never happen again!
After everything was wired up
with the correct pinout for the sonar
module (eight-pin vs. nine-pin), I ran
one of the test routines in the Handy
Board Libraries called sonar_display().
The initial test was done with a single
transducer wired up directly to the
ranger module. It is always best to test
each section and add a little at a time
so it is easier to troubleshoot any issues
that may (okay, will) come up.
When the tests ran, I could hear the
sonar transducer clicking and sending
out the pings like it should but I wasnt
getting any valid data back on the LCD
screen. It is set up so that if it doesnt
get a reading or it is out of range, then
it shows ***** as the distance.
Some troubleshooting was in order.
Was the Polaroid sonar board bad? Was
the wiring/pinout correct for the module?
Or was it some other problem? I knew it
must be close or the module wouldnt
have been firing the sonar at all.
SERVO 03.2008 47
How boards look inside the robot.
Pin Number
Nine-pin Module
(new style)
Eight-pin Module
(older style)
1 GND GND
2 BLNK BLNK
3 BINH
4 INIT INIT
5 FIL FIL
6 OSC OSC
7 ECHO ECHO
8 BINH V+
9 V+ N.A.
TABLE 1
As it turns out, the issue was a silly
one. I had left off an extra wire on the
expansion board. Since I had never
used the sonar section with a Handy
Board before, this was never an issue.
The line used to listen for the return sig-
nal is the only one that doesnt directly
plug in when you plug in the expansion
board! (That is what that extra wire is
for!) I just added that jumper and then
the sonar started giving valid readings.
Now that it was confirmed that
the Polaroid sonar ranging was work-
ing, the next step was to connect it to
the relay based board to multiplex the
transducers. It connects with a two-pin
.100 inch header connector. The con-
nection from the sonar MUX board to
the transducers is a bit unique.
BOB had four of the five transduc-
ers wired up. The wire used is the thin
RG-174 coax for each sensor. The fifth
sensor (which points down) wasnt
connected. The four sensors that were
wired have individual signal leads and a
common ground that went to a six-pin
(2x3) connector using five of the six
available pins. This connector appeared
to be one made by Molex but it is an
odd one. I could only find the matching
connector for it at Digi-Key, but I had to
get at least 10 if I wanted them.
Since it was an odd one that I
doubt I would ever need again, it was
hard to justify the purchase. Instead, I
noticed that a regular Molex female
pin would fit perfectly on the male pin
in that connector. So, I just crimped a
female pin directly on each signal lead
and on the common ground. I then put
heat shrink on each pin and instead of
having a shell, each one just plugs in
directly. To keep them in order, I used
an ohmmeter to confirm which lead
was for each transducer.
As they were identified, I used a
silver Sharpie pen to draw a number of
lines matching the number I gave to
each sensor. They fit snugly and a
larger piece of heat shrink was added
around the whole group to keep the
pins from coming out. Although there
was an open spot in the original con-
nector with a pin in it that could have
been repurposed, I left it alone since it
appears they never used it. Instead, I
just ran a new piece of RG-174 coax
from the fifth sensor, pointing down,
to a new separate connecter that was
added on the sonar MUX board.
Once connected, I ran the sonar
demo again for each sensor to make
sure each one was working. This part
worked out remarkably well! All the
sensors did their job and functioned
perfectly. I really liked having the LED
at each relay to show which one was
selected and it looks cool, too. Some
people feel that the fifth sonar which
faces down is pretty useless but I have
found it does have some value after all.
It is great to determine if someone
places a hand right in front of the
mouth area. A quick sample program
checked the fifth sensor and if a hand
was placed in front of it, BOB would
start making beeping sounds like R2-
D2. It was quite a good feeling to see
all those old sensors working again.
Getting feedback from
the drive motors (or not)
Being able to move the main drive
motors is nice, but it would certainly be
better to get some feedback from them
to make sure they do as they are told.
Without any feedback, all we can do is
turn on each motor, wait while it
moves, and hope that it actually did
move. Another problem is that since
motor power is usually derived directly
from the battery/power source, their
speed may slow down as the batteries
drain and hence the distance traveled
will decrease if you are only basing your
movements on fixed amounts of time.
Its probably best in this case to try
and get some feedback from the
robots motors/wheels to get an idea
of how far it has traveled and if the
wheels really moved. This is certainly
not going to be the only method to let
the robot know its moved around.
There are other factors to consider
(wheel slippage, etc.) that may prevent
the robot from reaching where you
think it has, based on the encoder
readings. These should really be used
as a guide to get you in the area and
later use a mix of other sensors to help
validate you are where you should be.
The Androbot base already has an
encoder on each drive motor that pro-
vides quadrature encoding. This provides
count and direction of movement. Just
four leads went to the encoder board:
+5V, Ground, Channel A, and Channel B.
The Handy Board includes several
add-on library routines. One in particu-
lar qencoders.icb will keep track of
a pair of quadrature encoders for you. It
watches digital inputs 10,11 and 12,13
to read the status of each encoder.
I loaded this up and tried it out.
Nothing happened. Since this was the
first time trying out some code and the
state of the old encoders was unknown,
I had to review all the parts. I started with
the encoders themselves. To test these, I
just made an adapter for my existing H2K
arm tester device. It is a very handy little
tool that can move the motor and show
the status of a couple encoders. With
that, I could determine that the encoder
itself was functioning okay.
After some more detective work, I
Poorly crimped connections.
48 SERVO 03.2008
From looking at the TOPO robots, Im
pretty sure that the Androbot engineers
were aware that as the batteries drained, the
speed of the main drive motors (and
distance traveled per unit time) would
decrease. This occurs because the motors
are powered directly off the battery and
there isnt any velocity feedback. Only the
original TOPO robot was open loop and
used the time the motor was on to try and
position the robot. The power for the drives
on that early TOPO robot came from a heavy
duty 18V regulator so the voltage would
remain a steady 18V even as the batteries
drooped. The power board was calibrated
so that the power on the robot would shut
off before getting low enough to prevent that
18V regulator from working. The later TOPO
II, III, and BOB prototypes did have feedback
from the motors so they could run right at a
full 24V. When those motors slowed down,
the controller could compensate.
NOTE
found the problem. It was a bad cable
assembly. This happens more that you
might think. It was a poor automated
crimp connection where the machine had
stripped the insulation off, but the wire
pushed in and the pin was crimped on the
insulation and not on the wire itself. It was
an open connection! (Didnt something
like that bad connection show up in
the article I wrote a while back on ARTI?)
I crimped on new pins and put it all
back together. Now, I started to see real
encoder values that would go up/down as
the wheel moved. Id also see the velocity
being calculated by the qencoders.icb driv-
er as well. Sweet! The encoders worked
and were getting data. Unfortunately, the
data ended up being pretty strange!
During the initial bench testing, the
motors werent running at full speed. The
next step was to watch the encoders
while ramping up the speed, wait a bit,
then ramp the speed back down to zero.
That is when things got interesting. A
very peculiar thing started happening.
The encoder and velocity started to rise as
expected, but as the motors sped up, the
counts started counting backwards! As it
sped up some more, it counted forward
again. This cycle continued back and forth
a bit and then the counts were erratic.
It appears that the encoder was
exceeding the polling time used and
the encoder library would start missing
pulses as the motor sped up. For the
moment, I am just going to use these
encoders for controlled closed loop
movements at low speed. Eventually, I
will either add a set of reflective encoders
right over the wheels or scale the value of
the encoder before it reaches the Handy
Board. Lower resolution of the encoder is
certainly better than completely unreli-
able values. I may lose a bit of encoder
resolution that way, but it should still be
close enough for this particular project.
Conclusion
A lot of ground was covered this
time. How to mount all the different
robot building blocks which include the
main Handy Board controller, H-bridge,
power distribution, and sonar boards. An
example on how the power distribution
and charging system can be done was
shown. A custom board was made using
come CMOS logic ICs and some DIP relays
to multiplex a Polaroid sonar module so
that it can drive the five sonar transducers
on BOBs head. Just like most of my
projects, all this just bolts and plugs right
in. You cant get any easier than that!
Now the Handy Board can move BOB
around under his own power and
sense some of the world around
him with his sonar sensors. Power is
available for the existing and future
circuits. The next topics that are
going to be covered are adding a
voice using the SpeakJet chip,
adding a co-processor to control
the head and lights, getting better
data from the encoders on the
drive system, and connecting a few more
sensors so that BOB can sense and
interact more with the world around him.
To help wrap up this project (like
they are ever really finished), some pro-
gram examples will be provided in the
next article, as well as some additional
building blocks of code to make future
programming of BOB much easier.
Keep those old robots alive! SV
Robert Doerr can be reached via email
at rdoerr@bizserve.com.
Original drive motor encoder.
SERVO 03.2008 49
RobotWorkshop Creator of the
plug-in H-bridge boards for the L293D
www.robotworkshop.com
Main Handy Board site
(also Handy Cricket links)
www.handyboard.com
Information on the Handy Board
Expansion board
www.handyboard.com/hbexp30
Tutorial on Polaroid sonar modules
on Acroname website
www.acroname.com/robotics/info/
articles/sonar/sonar.html
Acroname Supplier of SensComp
6500 (new Polaroid driver board)
www.acroname.com/robotics/parts/
R11-6500.html
Supplier of Polaroid sonar
ranging modules
www.1sorc.com
Note from Toshiba about CMOS logic
www.semicon.toshiba.co.jp/eng/
product/logic/faq/answer08.html
Site about modification and use
of Polaroid modules scavenged
from cameras
www.uoxray.uoregon.edu/polamod
Article by Dennis Clark on
Polaroid sonar modules
www.seattlerobotics.org/encoder/
200010/dlcsonar.html
Datasheet on DC/DC power module
www.cincon.com/pdf/EC7A.PDF
WEB REFERENCES
// Routine to energize appropriate relay on
// Sonar MUX board to select desired sonar
// (transducer number 0-4). Any value 4 or
// larger automatically selects the mount
// sonar.
void sonar_select(int rly) {
if (rly==0) {
clear_digital_out(0);
return;
}
if (rly==1) {
set_digital_out(0);
return;
}
if (rly==2) {
set_digital_out(1);
return;
}
if (rly==3) {
set_digital_out(0);
set_digital_out(1);
return;
}
else {
set_digital_out(2);
}
}
Subroutine to Demonstrate
Sonar Select MUX Board
50 SERVO 03.2008
N
ow that I have a Minebea
29SM-K series 1.8-per-step
stepper motor shaft spinning
back and forth on the bench, I figured
it might be a good time to sit down
and tell you how I persuaded an
STMicroelectronics L6208PD to push
current through the Minebeas stepper
motor coils. Unfortunately, I cant provide
a visual play-by-play of my Minebeas
alternately spinning motor shaft.
However, I can give you a description of
the hardware and software stuff you will
need to make your motor spin under the
control of an STMicroelectronics L6208PD
and a PIC18F2620 microcontroller.
Over the years, Ive read thousands
of datasheets and application notes. To
this day, it is still a thrill to translate
the words of the manufacturers into a
functional device. In the previous SERVO
L6208PD installment, we walked around
the L6208PD with the intent of putting
the L6208PD to work with a stepper
motor and PIC microcontroller. All of the
preflight work we expended last month
will be applied to our goal this month of
creating controlled rotational movement
of a stepper motor shaft.
Spinning the Printed
Circuit Board
The beginnings of our L6208PD
Motor Driver Board can be seen in
Screenshot 1. Ive used ExpressPCB as
the printed circuit board design and
layout program so that you can either
copy my printed circuit board layout or
use my original printed circuit board pat-
tern as a basis for your own L6208PD
printed circuit board (PCB) design.
Note that the L6208PD Motor
Driver Board design provides a relatively
large heatsink area for the L6208PD on
the top side of the PCB. Its a good idea
to physically bond the L6208PDs
heatsink slug to the heatsink pad area.
So, as you can see in Screenshot 1, Ive
included some top-to-bottom thru holes
under the L6208PDs heatsink lug to
provide solder access once the L6208PD
has been mounted on its pads.
Before mounting the L6208PD, I
applied some solder paste to the
L6208PDs heatsink lug. After complete-
ly soldering down all of the L6208PD
pins, I used a high-wattage soldering
iron applied to the bottom side heatsink
pad to flow the solder paste on the
L6208PD heatsink lug. I can tell you from
practical experience that this method of
mounting the L6208PD is sound.
While experimenting with the
by Fred Eady
Building a
STEPPER MOTOR
CONTROLLER Part 2
SCREENSHOT 1. The idea is to provide as much heatsink area as possible for the
L6208PD. We also have to adhere to the L6208PD printed circuit board guidelines to
keep electrical noise to a minimum.
Eady2.qxd 2/5/2008 6:19 PM Page 50
L6208PD software, I managed to over-
heat and damage the L6208PD. I used
a standard soldering iron to disengage
each of the L6208PDs 36 pins and the
high-wattage soldering iron to break
the solder bond between the PCB
heatsink area and the L6208PD heatsink
lug. The L6208PD removal operation
took less than 10 minutes to perform.
The L6208PD pin layout is very well
thought out. All of the L6208PD ground
connections lie on the four corner pins
of the device. The L6208PD heatsink lug
is also attached internally to the devices
grounding system. Thus, all we have to
do is pull our combination ground/
heatsink pad through the center line of
the L6208PD pads, attaching the four cor-
ner L6208PD ground pins in the process.
The L6208PD datasheet recom-
mends that a star ground point be
placed at the intersection of the
ground return point of the H-bridge
sense resistors and the bulk power
supply ground entry point. A 100 F
bulk filter capacitor (C3 in the schemat-
ic) is also located at this star ground
point. The star ground point serves as
the single-point connector area for the
signal and power grounds, as well.
Large pad areas are used to
connect the H-bridge current sense
resistors (R9 and R10) between the
L6208PD SENSEA/SENSEB pins and the
star ground return point. Take another
look at Screenshot 1 and you will see
that the same large pad area design
point is also applied to the bulk
power (VSD, GND) and motor winding
terminations (1A, 2A, 1B, 2B).
A pair of 0.041 inch pads with
0.020 holes are used to interface the
L6208PDs H-bridge output pins to their
termination points. The thru hole pads
provide a high-current path from the
L6208PD H-bridge pins on the top side of
the PCB to the H-bridge termination pins
that are fed from large pad areas on the
bottom side of the PCB. Note also that
the incoming bulk supply voltage is dis-
tributed to the L6208PD power pins using
a large bottom side pad and 0.031 inch
thru holes to the L6208PD power pins.
I really like using screw terminals for
motor project boards. However, it can be
a pain in the rotor to screw and unscrew
motor leads and power leads when you
need to fix or make changes to
the boards electronic components.
The ideal solution is to design in
removable screw terminations.
For this project, I used 5 mm
pitch PCB mounted terminal block
headers. The terminal block head-
ers accept 5 mm removable screw
terminal blocks. The removable
screw terminal blocks I used come
as an assembly of two screw
terminations on a removable block.
The terminal block pairs can be
coupled to other terminal blocks to
form a single integrated terminal
block of screw terminals that can be
removed and mounted as a single entity.
The 5 mm terminal block headers
and their matching 5 mm screw terminal
block assemblies can be had from
Digi-Key. The 24-pin breakable 5 mm
terminal block part number is ED1682-
ND. The two-position 5 mm terminal
block plug assembly can be ordered from
Digi-Key as part number ED1613-ND.
Since you cant see the 5 mm termi-
nal block assemblies completely when
they are mounted, Ive provided a shot of
the terminal block headers and the termi-
nal block plug-in assemblies I used in this
project in Photo 1. You can also get the
terminal block plugs with 3, 4, 5, 6, 7,
10, or 12 screw terminals per assembly.
Now that the foundation of the
L6208PD Motor Driver Board is in
place, lets flesh out the supporting
hardware. All of the L6208PDs
companion componentry is laid down
on our driver board in Screenshot 2.
My L6208PD application requires a
bulk voltage between 12 and 24 volts.
Thus, the voltage rating for C3, in my
case, can be safely set at 35 volts. The
L6208PD can accept bulk voltages up
to 52 volts. So, choose your working
voltage for C3 accordingly.
C1, C2, and C9 are all rated for
100V. C1 and C2 are standard ceramic
chip capacitors in a 1206 SMT package. It
is recommended that C9 be a higher qual-
ity type of capacitor. My C9 is a thru hole
mounted 0.22 F metalized polyester film
capacitor (Digi-Key BC1647-ND).
H-bridge sense resistors R9 and R10
should be non-inductive, which eliminates
SERVO 03.2008 51
PHOTO 1. Isnt it funny? Here we are designing
a device around a piece of high-tech silicon
and the least complex piece of the puzzle a
screw terminal demands our attention.
SCREENSHOT 2. Things are starting to shape up and make sense. With the L6208PD
supporting hardware in place on the printed circuit board, all we have to do is run
some traces to put the capacitors and resistors to work for us.
Eady2.qxd 2/5/2008 6:19 PM Page 51
52 SERVO 03.2008
any power resistor type that is composed
of a piece of resistive wire wound in a
spiral inside of its package. I chose the
Vishay/Dale WSL2512 series. These
resistors are rated for one watt and have
a tolerance of 1%. I did not experience
any noticeable heating of the WSL2512
sense resistors and the L6208PD
datasheet calls out one watt resistors.
However, if you choose to, you can put a
two watt resistor in the R9 and R10 seats.
If you go with the two watt
resistor (WSL2816), youll need to
rework the PCBs H-bridge sense
resistor pads to fit the slightly larger
SMT package. (The Digi-Key part
number of the one watt WSL2512
resistors I used is WSLG-.20CT-ND.)
A pair of 1N4148 switching diodes
can be used in place of the integrated
diode pair labeled as D1. In addition to
being more compact, the single SOT-323
diode package used here is easier to han-
dle. The BAT46SW device is a pair of
back-to-back diodes that are rated
individually with a breakdown voltage of
100V and a maximum forward current of
150 mA. Recall from our discussion last
month that D1, C1, R1, and C9 make up
a charge pump that drives the high side
MOSFETs of the L6208PDs H-bridges.
The L6208PD charge pump will
provide a voltage to the H-bridge
MOSFET gates that is about 10 volts
higher than the incoming bulk voltage.
Even at the maximum bulk input
voltage of 52 volts, the 100V rating on
the charge pump capacitors is more
than adequate for any application.
Resistors R3, R4, and capacitor C5
form a combination voltage divider/low-
pass filter for input VREFA. Capacitor C6,
along with resistors R5 and R6, perform
the same function for the L6208PDs
VREFB input. The output current of the
H-bridge pair is directly affected by the
voltages at the VREFA and VREFB inputs.
So, with our setup we can provide a
steady state DC voltage to the VREF
inputs or use a PWM or pair of PWMs to
supply the current reference voltages.
The L6208PDs H-bridges can be
disabled by applying a logical low level
to the R2-C4 RC pair. Recall that the RC
time constant provided by R2 and C4
provides a recovery off-time in the
event of an overcurrent condition.
While we have off-time on our
minds, the RC pairs of R7-C7 and R8-C8
determine the off-time of the PWM
current controller monostable. Theres
no science behind the inclusion of the
R7 and R8 resistors as potentiometers.
I placed 100K potentiometers in the R7
and R8 positions to allow for easy
tinkering. I found that I could play a
part in controlling the motor current
and thus its operation by adjusting the
off-time of the PWM current controller
monostable using a screwdriver.
With the layout for the L6208PD and
its band of brothers down on fiberglass,
lets turn our attention to planting some
pads for the PIC. Obviously, applying the
maximum VSD voltage to the PIC VDD
pin will allow the Microchip magic smoke
to escape. So, well need to design in a
voltage regulator system that can
withstand the 24 VDC maximum bulk
voltage at its input and provide +5.12
VDC to the VDD pin of our PIC18F2620.
Since there are no fixed linear
voltage regulators that will produce
exactly +5.12 VDC at their outputs,
well resort to designing in a variable
voltage regulator. I have chosen
the low-dropout first cousin of the
venerable LM317, the LM1084-ADJ.
If youre wondering why I say we
need to have +5.12 VDC on the
PIC18F2620s VDD pin, the answer is
analog-to-digital converter. We wont
be employing the services of the
PIC18F2620s 10-bit analog-to-digital
converter subsystem in this project.
However, there is a possibility that you
may want to use the PIC18F2620s A-to-
D converter in your L6208PD project.
Were using the LM1084-ADJ because it
regulates effectively with input/output
voltage differentials of up to 29V. Sure,
we could have used the fixed LM1084-
5.0 to give us a nominal +5 VDC for the
PIC as it has a 25V input/output voltage
differential. However, saving that pair of
sub-penny resistors is a solution I
refused in favor of a more accurate A-to-
D converter reference voltage.
Using the PIC18F2620s VDD input
voltage as the A-to-D converter
reference voltage yields a 5 mV per
step resolution on your A-to-D converter
readings when VDD is equal to +5.12
volts. Regardless of which LM1084
logic power supply solution you choose
to use in your project, you cant exceed
35V at the VSD input with our LM1084
power supply design. If you exceed 30V
at the VSD input with the LM1084-5.0
or 35V at the VSD input with the
LM1084-ADJ, youll have to supply the
+5 VDC logic power in another way.
What you see on the fiberglass in
Screenshot 3 and in the power supply
section of the L6208PD Motor Driver
Board schematic works well for this par-
ticular spin of the L6208PD electronics.
Schematic 1 shows a variable
value for the voltage adjust resistor R11.
SCREENSHOT 3. A fixed voltage regulator would work well here. Again, I have provided
a tinker knob for those of you that live to tinker.
Eady2.qxd 2/5/2008 6:20 PM Page 52
+
5
V
D
C
+
5
V
D
C
+
5
V
D
C
+
5
V
D
C
V
S
D
+
5
.
1
2
V
D
C
V
S
D
V
R
1
L
M
1
0
8
4
I
T
-
A
D
J
3
1
2
V
I
N
ADJ
V
O
U
T
C
1
4
.
1
u
F
C
1
5
.
1
u
F
R
6
1
5
K
+
C
3
1
0
0
u
F
R
5
5
6
K
C
2
.
2
2
u
F
C
9
.
2
2
u
F
C
1
6
.
1
u
F
R
8
3
9
K
C
6
.
0
1
u
F
R
1
1
0
0
C
8
.
0
6
8
u
F
B
I
P
O
L
A
R

S
T
E
P
P
E
R

M
O
T
O
R
R
7
3
9
K
R
4
1
5
K
+
C
1
0
1
0
0
u
F
R
1
5
1
0
K
U
3
S
T
3
2
3
2
1
3 8
1
0
1
1 1 3
4 5
2
6 1
2
9 1
4
7 1
5
1
6
R
1
I
N
R
2
I
N
T
2
I
N
T
1
I
N
C
1
+
C
1
-
C
2
+
C
2
-
V
+
V
-
R
1
O
U
T
R
2
O
U
T
T
1
O
U
T
T
2
O
U
T
V
S
S
V
D
D
C
1
7
.
1
u
F
C
4
5
6
0
0
p
F
O
P
T
I
O
N
A
L
A
C
T
I
V
I
T
Y

L
E
D
O
P
T
I
O
N
A
L
3
3
0
R
1
1
V
A
R
I
A
B
L
E
C
1
2
.
1
u
F
R
9
0
.
2
0
+
C
1
3
2
2
0
u
F
D
1
B
A
T
4
6
S
W
R
1
2
1
2
0
I
C
S
P

C
O
N
N
E
C
T
O
R
1 2 3
4 5 6
1 2 3
4 5 6
C
1
1
.
1
u
F
C
1
.
0
1
u
F
R
1
3
1
0
0
R
2
1
0
0
K
U
2
P
I
C
1
8
F
2
6
2
0
2 3 4 5 2
1
2
2
2
3
2
4
2
5
2
6
2
7
2
8
1
1
1
2
1
3
1
4
1
5
1
6
1
7
1
8
1
0
9
1
6 7
8
1
9
2
0
R
A
0
R
A
1
R
A
2
R
A
3
R
B
0
R
B
1
R
B
2
R
B
3
R
B
4
R
B
5
R
B
6
/
P
G
C
R
B
7
/
P
G
D
R
C
0
R
C
1
/
C
C
P
2
R
C
2
/
C
C
P
1
R
C
3
R
C
4
R
C
5
R
C
6
/
T
X
R
C
7
/
R
X
O
S
C
2
/
R
A
6
O
S
C
1
/
R
A
7
M
C
L
R
R
A
4
/
T
0
C
K
I
R
A
5
G
N
D
G
N
D
V
D
D
U
1
L
6
2
0
8
P
D
1
36
18
19
2
7
12
1
1
1
5
2
6
1
3 9
2
4
30
5
25
33
4
2
8 8
3
2
1
0
7
2
2
2
9
GND
GND
GND
GND
H
A
L
F
/
F
U
L
L
SENSEA
C
W
/
C
C
W
O
U
T
1
A
V
R
E
F
B
R
C
A
V
R
E
F
A
R
C
B
VBOOT
O
U
T
2
A
SENSEB
VSB
VSA
C
O
N
T
R
O
L
R
E
S
E
T
O
U
T
2
B
C
L
O
C
K
VCP
O
U
T
1
B
E
N
C
1
8
.
0
4
7
u
F
C
5
.
0
1
u
F
C
2
1
.
3
3
u
F
R
1
4
1
K
C
1
9
.
3
3
u
F
C
2
0
.
3
3
u
F
R
1
0
0
.
2
0
R
3
5
6
K
D
B
9

F
E
M
A
L
E
1
2
3
4
5
6
7
8
9
1
2
3
4
5
6
7
8
9
C
7
.
0
6
8
u
F
SERVO 03.2008 53
SCHEMATIC 1. The L6208PD is
easy to support and so is the
PIC18F2620. If you dont need to
talk to the outside world, you can
eliminate the RS-232 circuitry.
Eady2.qxd 2/5/2008 6:20 PM Page 53
54 SERVO 03.2008
Lets use the following LM1084-ADJ
output voltage formula to see what we
would have to plug in for R11 to get
+5.12V at the LM1084-ADJ output pin:
V
OUT
= 1.25V(1 + (R2/R1))
5.12V = 1.25V(1 + (R2/120))
R2 = 371.52
The closest 1% value in a 0805
package is 374. Lets see what that
will get us:
V
OUT
= 1.25V(1 + (R2/R1))
V
OUT
= 1.25V(1 + (374/120))
V
OUT
= 5.1458V
The resultant voltage of +5.1458V is
too high. Okay, now lets go the other way
and select the next lower value of 365:
V
OUT
= 1.25V(1 + (R2/R1))
V
OUT
= 1.25V(1 +
(365/120))
V
OUT
= 5.0521V
I chose to dial in +5.12V using a 5K
potentiometer soldered across the R11
pads. However, if the accuracy of the A-to-
D converter is of no concern to you, you
can drop in a 365 1% 0805 SMT chip
resistor for R11 and live happily ever after.
Now that weve gone to the trouble
of all of that thinking on the logic power
supply voltage level, lets put a load on the
crazy thing. As you can see in Screenshot
4, Ive added the PIC18F2620, the
ICSP portal, and an RS-232 subsystem
based on an STMicroelectronics
ST3232. Ive also connected
the dots between all of the
electronic glue components,
U1, U2, and U3. Believe it or
not, there were no cockpit
errors in the first run of this PCB
and I actually got to use with-
out any modifications the
first spin PCB I received from
the folks at ExpressPCB.
Proving the Concept
Yes, there are things
missing in Photo 2. I went
ahead and put down all of the
passive components and the L6208PD
charge pump diodes as I used a batch
oven to reflow most of the mounted com-
ponentry you see in Photo 2. After the
reflow process, I hand-soldered the 5K
voltage adjust potentiometer, the ICSP RJ-
12 jack, the nine-pin female shell connec-
tor, and the 5 mm VSD terminal header
onto the L6208PD Motor Driver Board. I
purposely omitted the L6208PD from the
soldering process as I have some things to
do before mounting the L6208PD on its
pads and soldering its heatsink lug to the
upper and lower PCB heatsink pads.
The first order of business involved
is applying a very low Design Cycle volt-
age to the VSD header. Before applying
any voltage to the VSD header, I made
sure that the voltage adjust potentiome-
ter showed a ground (zero volts) to the
LM1084-ADJs adjust pin. Apply a value
of zero to R2 in the LM1084-ADJ output
voltage formula we discussed earlier and
youll get 1.25V as your output voltage.
Using a variable power supply that
indicates the current draw, I slowly
increased the voltage at the VSD head-
er to a point that provided a steady
1.25V on the output of the LM1084-
ADJ. I then ran the VSD voltage up past
9 VDC looking for excessive current
draw and signs of escaping magic
smoke. The current draw was minimal
and I didnt have to light up any incense
sticks. With a green light for the cur-
rently mounted circuitry, I adjusted the
5K potentiometer to give me +5.12
VDC at the LM1084-ADJs output.
Recall that earlier I stated that we
could use a PWM signal to provide our
reference voltages for the L6208PDs
VREF inputs. A quick look at the L6208PD
schematic shows that I have connected
the pair of PIC18F2620 PWM outputs
to the pair of low-pass filters that are
buffering the L6208PD VREFA and VREFB
inputs. Ive also assigned and connected
PIC18F2620 I/O pins to the L6208PD
control pins (HALF/FULL, CW/CCW,
CONTROL, RESET, CLOCK, AND EN).
For now, all Im interested in doing
is generating a sine and cosine wave at
the L6208PD end of the low-pass filters.
Once I have the sine and cosine wave-
forms, Ill need to double toggle the
CLOCK line at every 90 transition point
of the sine wave. Ill explain more about
SCREENSHOT 4. Here she is! Ill supply this completed L6208PD Motor Driver Board
ExpressPCB file for you by way of the SERVO website (www.servomagazine.com).
PHOTO 2. With no prior experience with the
L6208PD, I didnt want to take a chance on letting
the STMicroelectronics magic smoke escape due to
something silly I may have missed in the design. So,
Ill prove the concept and adjust the voltages
before I weld the L6208PD into place.
Eady2.qxd 2/5/2008 6:21 PM Page 54
the double-tap clock signals later on.
Lets generate a sine table that
represents a full step of 90 as 16
microsteps. Screenshot 5 is a view of
an Excel spreadsheet that I whipped up
to do the sine-to-PWM value conver-
sions. The Excel sine-to-PWM duty cycle
formula for an eight-bit PWM unit is:
PWM Duty Cycle Value =
SIN(RADIANS(ACELL_VALUE)) * 255
where ACELL_VALUE is the value of the
DEGREES PER STEP cell in column A;
255 is the eight-bit sine table multiplier.
Excel uses RADIANS instead of
DEGREES in its sine calculations. The
RADIANS statement in the Excel
formula converts the RADIANS value to
DEGREES in the computation. Since we
are computing the PWM duty cycle
value for an eight-bit PWM subsystem,
we use eight bits (0xFF or 255) as our
sine table multiplier. The PIC18F2620
contains a 10-bit PWM subsystem,
which allows us to use a 10-bit sine table
multiplier of 1023 (0x3FF). The results of
the 10-bit PWM duty cycle calculations
are also shown in Screenshot 5.
Allow me to illustrate why we only
need to compute and load the 16 PWM
duty cycle values. The L6208PD will
settle into a home position of 45 fol-
lowing power-up or a PIC18F2620 initi-
ated reset operation. This 45 position
is not a physical location of the stepper
motors rotor shaft. It is a full step posi-
tion state reflected in the L6208PDs
step generator and phase generator.
If you pull out your scientific calcu-
lator and take the sine and cosine of
45, you will find that the result for both
is 0.707 and some change. In stepper
motor land, that means that the current
though the coils which are, in this
case, aligned to a full step home posi-
tion is equal. The L6208PDs internal
phase generator is also sitting at its
home position along with the L6208PD
step generator. At this point, the PWM
duty cycle value associated with the 45
state has been loaded into the
PIC18F2620s CCPR1L and CCPR2L duty
cycle registers by our firmware.
Lets start our sine and cosine
cycles by moving through the 16 sine
entries. Remember, a cosine wave is
90 out of phase with a sine wave. We
will move one of our PWM duty cycle
registers from 45 to 90 and then back
through 45 to 0 and up again to 45.
Simultaneously, we will move the
other PWM duty cycle register from
45 to 0 and then back up through
45 to 90 and down again to 45. If
you were to graph this out by hand,
you would come up with something
similar to the CleverScope capture Ive
snared in Screenshot 6.
By simply rotating through the PWM
duty cycle sine table values with a 90
phase difference, we can use a pair of
PIC18F2620 PWM outputs to generate
the necessary sine and cosine waveforms
we will need to microstep our Minebea
stepper motor using the L6208PD
mounted on our Motor Driver Board.
I switched the CleverScope scope
probe off of the second PWM and onto
the L6208PD CLOCK pin. To properly
microstep with the L6208PD, we must
generate a clock pulse on every required
phase reversal. In other words, a clock
pulse must be generated every time we
cross a 90 boundary in the sine table.
SERVO 03.2008 55
SCREENSHOT 5. This is all we need to generate sine and cosine functions for our L6208PD.
All we really care about is putting together 90 phase transitions with the PWM. The
L6208PDs built-in phase generator takes care of the motor state transitions for us.
SCREENSHOT 6. This is how the CleverScope saw my PWM-generated sine and cosine
signals. Note that you can make out the PWM steps inside of the sine and cosine
waveforms. There are 16 PWM steps for each 90 phase.
Eady2.qxd 2/5/2008 6:21 PM Page 55
The CleverScope screen capture I
collected in Screenshot 7 verifies that my
proof of concept firmware is on the mark.
The double toggle of the CLOCK line is a
trick that fools the L6208PD into disabling
the H-bridge (taking the current to zero)
at the appropriate time during the
microstepping process. The L6208PD
must be configured for HALF step mode
to make this trick sell. Heres how it works.
When the desired current is set to
zero (every 90 transition), a clock pulse
is applied to the L6208PD and the last
PWM duty cycle table value for that 90
phase is loaded. A second clock pulse is
then sent to the L6208PD and the first
PWM duty cycle microstep value for the
next 90 phase is loaded into the PWM
duty cycle register. The L6208PDs step
sequence generator will initiate a chain
of events within the H-bridge as a result
of the double clock sequence. Current
flowing in one direction in the H-bridge
will be halted as the H-bridge will
become disabled. Then, the current in
the affected H-bridge will begin to flow
in the opposite direction.
For instance, lets begin at the
bottom of one of the sine waveforms
shown in Screenshot 7. Our desired
current is zero at this point. So, we fire
off a clock pulse and load our last PWM
duty cycle value for that 90 run. We
then fire off yet another clock pulse and
load up our first PWM duty cycle
microstep value for the next 90 run. We
run the full 90 in the sine table uphill
and repeat the process to start going
downhill on the sine waveform. We get
to the bottom of the hill and do it all
over again. Get the idea? Were simply
fooling the L6208PDs phase generator
and step generator to sync up with our
sine table values at phase reversal times.
It should be obvious that if you
dont use the PIC18F2620s PWMs for
microstepping, you can simply set a
PWM duty cycle level and hold there to
provide a voltage reference for the
L6208PDs VREFA and VREFB inputs.
Once youve established your desired
voltage level which translates to the
desired motor current level you can
sit there and spin the stepper motor
rotor with a series of clock pulses.
A Walk in the Park
With the questions erased as to how
to generate microstepping PWM
waveforms with interlocking clock
pulses with a PIC18F2620, I tacked the
L6208PD into place on the Motor Driver
Board and installed all of the remaining
electrical components. The completed
and fully functional L6208PD Motor
Driver Board is shown in Photo 3.
The code behind all of the sine and
cosine waveforms youve been looking at
is actually very simple. It all begins with
transposing those Excel spreadsheet
numbers into a pair of sine tables in the
PIC18F2620s program memory area. The
sine and cosine tables are actually simple
arrays of 16-bit words. Heres what the
tables look like in source form:
unsigned int const sinetable1[] = {723,
790,850,902,945,978,1002,1018,102
3,1018,1002,978,945,902,
850,790,723,648,568,482,
391,296,199,100,0,100,199,
296,391,482,568,648};
unsigned int const sinetable2[]
= {723,648,568,482,391,296,
199,100,0,100,199,296,391,
482,568,648,723,790,850,
902, 945, 978, 1002, 1018,
1023,1018,1002,978,945,902,
850,790};
The 45 position in
sinetable1 is at the first 723 mark
followed by the 90 position at
SCREENSHOT 7. This is an expanded view of the double clock pulses relative to the 90
phase reversal points of the PWM-generated microstepping sine waveform. The sine
wave shape is clear in this shot and you can almost count all 16 of the PWM voltage
steps that make up each 90 phase.
PHOTO 3. The L6208PD package looks tight. However,
I had absolutely no trouble hand-soldering it onto
the printed circuit board. After burning the H-
bridges, I had little trouble in removing the dead
L6208PD and installing a fresh one in its place.
56 SERVO 03.2008
STMicroelectronics (www.stmicro.com)
L6208PD; ST3232
Microchip (www.microchip.com)
PIC18F2620
Saelig Corporation (www.saelig.com)
CleverScope
Express PCB (www.expresspcb.com)
PCB design and layout program
The L6208PD Motor Driver Board
code was written with the HI-TECH
PICC-18 C compiler.
Sources
Eady2.qxd 2/5/2008 6:22 PM Page 56
the 1023 mark. The 135 position begins
at the second 723 marker with the 0 posi-
tion being marked as named. Sinetable2
begins 90 out-of-phase with sinetable1
with the first 723 value representing 135.
As with sinetable1, the 0 position
is marked by a zero value in the table.
The second 723 value in sinetable2 is
the 45 position, which is followed by
the 90 position at the 1023 marker.
The sinetable1 and sinetable2 values
are continually rotated through from
beginning to end with a pair of clock
pulses occurring at every eighth and
ninth, 16th and 17th, 24th and 25th,
and 31st and 32nd value in the table.
Or, simply put, every 90.
The majority of the L6208PD step-
per motor driver code is simple on-off
I/O code. The PWM duty cycle loader
code is the most busy. Once the PWMs
are initialized, the 10-bit sinetable values
are parsed with the two least significant
bits being placed in the respective
CCPxCON register (CCP1CON or
CCP2CON) and the eight most signifi-
cant bits being routed to their respective
duty cycle register (CCPR1L or CCRR2L):
pwm_lsbs = (sinetable1[sinetable_
index] & 0x0003) << 4;
CCP2CON &= 0b11001111;
CCP2CON |= pwm_lsbs;
CCPR2L = sinetable1[sinetable_index]
>> 2;
pwm_lsbs = (sinetable2[sinetable_
index] & 0x0003) << 4;
CCP1CON &= 0b11001111;
CCP1CON |= pwm_lsbs;
CCPR1L = sinetable2[sinetable_index]
>> 2;
The PIC18F2620 holds the least sig-
nificant bits of the PWM duty cycle value
in bits 4 and 5 of the CCPxCON register.
The remaining eight bits are placed into
the CCPRxL duty cycle registers.
Microstepping Out
You now have all of the hardware
information you need (within the extents
of this article) to build up your own ver-
sion of the L6208PD Motor Driver Board
Ive presented. What you dont have is a
complete set of the HI-TECH C-based
L6208PD stepper motor driver source
code. Ill provide the C source to you via
a download from the SERVO website
(www.servomagazine.com). Ill also
post the ExpressPCB file and the Excel
sine table spreadsheet for your access.
And, by the way, you can complete-
ly assemble your L6208PD Motor Driver
Board and still play with the PIC and the
PWM stuff. Just set the L6208PD ENABLE
pin to zero in your initialization code.
That will disable the L6208PD internal H-
bridges so you can do your thing with the
motor control and PWM signals without
worrying about doing something bad to
your L6208PD. See you next time! SV
SERVO 03.2008 57
Fred Eady can be reached via email
at fred@edtp.com
Contact the Author
Let new ideas spring forth!
All modules shown actual size
TReX dual motor
controller
$99.95
TReX Jr: all the great features
of the full TReX in a lower-
power, lower-cost package.
TReX Jr
$59.95
Orangutan LV-168: full-featured
robot controller perfect for low-voltage
robots.ATmega168 uC, 2-channel H-
bridges, 8x2LCD, buzzer, and more.
Orangutan LV-168 $59.95
TRe X dual mot or cont rol l e r:
innovative design allows instant
switching between radio control
(RC) or analog voltage and
asynchronous serial
(RS-232 or TTL) with
a mu l t i t u d e o f
additional features.
Find out more at www.pololu.com or by calling 1-877-7-POLOLU.
Eady2.qxd 2/6/2008 3:33 PM Page 57
58 SERVO 03.2008
T
odays hobbyists can purchase robotic platforms with a wide variety
of electronic sensors, mechanical actuators, and programmable
microcontrollers. This means that constructing a powerful robot is
within the capacity of almost everyone, even those with a minimal
background in electronics. This availability of hardware is progressively
forcing robot enthusiasts to change their focus from building robots to
programming robots. As this emphasis on software grows, the number
of people interested in robotics will increase tremendously. This is
analogous to the boom that occurred in the personal computing industry
when people no longer had to design and build their own computers.
Just as talented programmers
shaped the personal computing
industry, the ever-increasing number
of robot programmers will carry
personal robotics to new levels as they
strive to create truly intelligent
machines. The artificial intelligence
(AI) algorithms needed for this
endeavor will be beyond the capabili-
ty of the microcontrollers used in
robots because they lack adequate
memory for data and are severely
limited in mathematical and array
manipulation capabilities. Additionally,
the limited debugging features of
microcontrollers are not suitable for
large, complicated AI projects.
You could, of course, embed an
entire PC in your robot, but this has
many obvious disadvantages. This
article will explore an easy and
powerful solution that counteracts the
deficiencies mentioned above, as
well as provide possibilities for further
fascinating innovations.
The Solution
If you have a PC program that
communicates wirelessly with a robots
onboard microcontroller, you can
eliminate the need for the robot to
carry complex computing power. In
such a configuration, the PC would
provide all the intelligence. The robots
controller would simply receive
commands from the PC and execute
them without regard for why the
actions have been requested.
The hardware-control program
running on the robot can be very
simple or moderately complex based
on your needs. Even a complex
hardware-control program though,
can be handled by most microcon-
trollers. Lets look at an example.
When the controller receives a
command from the PC to move
forward, it must actuate the motors
on the robot to do that. The actual
code implemented would depend
on the type of motors used by
your robot, as well as how they are
interfaced to the controller.
For simple applications, open-loop
control of the robots motors could
be very acceptable. If you need a
robot that can make precise
movements though, more complex
motor control routines might use
wheel encoders (or compasses and
other navigational devices) to ensure
that the requested motion has been
carried out accurately.
In addition to handling the
robots movement, the hardware-
control program must also acquire
and send back sensory data as
needed or requested by the AI
program running on the PC. The PC
will use this sensory data to decide
how to accomplish the high-level
goals. To achieve all this, we will need
a suitable programming language
for programming the PC to act as the
AI controller.
Blankenship.qxd 2/5/2008 6:14 PM Page 58
RobotBASIC
RobotBASIC is a free programming
language (www.RobotBASIC.com)
known for its integrated robot
simulator (see Figures 1 and 2) that
makes it easy to learn how to control
a mobile robot with a multitude of
sensory options. The latest version of
RobotBASIC supports parallel and serial
I/O ports, USB, and a variety of wireless
communications including Bluetooth.
These new features, along with all its
memory and mathematical capabilities,
make RobotBASIC very suitable as the
AI controller described above.
An innovative feature of Robot
BASIC is that many of the language
statements used to control the
simulated robot also have the ability to
automatically communicate with and
control a real robot. You can use
the simulator to develop complex
applications such as an algorithm to
navigate through a home or to locate
and utilize a battery charger. Once the
program is working properly on the
simulator, you can immediately use the
very same program to control a real
robot. The above two example applica-
tions and many others are explored in
detail in the book Robot Programmers
Bonanza that is due to be published in
mid-2008 by McGraw-Hill.
A Remote-Control Program
Lets look at an example that
illustrates the above ideas. Figure 3
shows a simple RobotBASIC program
that simulates a remote-controlled
robot. It allows the operator to press
the Up-Down-Left-Right arrow keys on
the keypad to make the
simulated robot move
around the screen.
The statement
rCommPort 0 (second
line in the main
program) is not actual-
ly needed during the
simulation, but will be used
later to switch RobotBASIC
from the simulation mode
to communicating with the
real robot.
At the top of the pro-
gram, some constants are
defined to represent the
codes for the arrow keys.
The main program sets the
foreground and back-
ground colors for the display and the
robot is initialized and placed at the
center of the screen. Finally, the
RemoteControl subroutine is called.
This subroutine enters in a loop
checking for a key press on the
keyboard. If any of the arrow keys are
pressed, the robot is moved according-
ly. The forward and backward move-
ments are only done if the bumper sen-
sors indicate that no objects block the
robots path. The program also displays
the status of the robots bumper sen-
sors at the top of the screen.
Imagine for a moment
that you have a real robot
with bumper sensors just like
the simulated robot. What if
you could tell the program in
Figure 3 to stop performing
the simulation and start con-
trolling the real robot? Whats
more, the real robot can use
nearly any microcontroller
and any motors. It can have
all of the sensors simulated in
RobotBASIC or just the few needed for a
particular project. The only requirement
is that the onboard microcontroller be
able to respond properly to commands
from RobotBASIC, according to the
protocol specified below.
The Protocol
Regardless of the robot command
involved, RobotBASIC will always send
SERVO 03.2008 59
FIGURE 1. This is an example
of the simulated robot
showing the correct path
through a maze. On the
first attempt, the robot tried
many deadends before
finding the solution.
FIGURE 3. Remote-
controlled robot simulation.
FIGURE 2. The graphic
capability of RobotBASIC
allows you to create
custom environments
such as this suite of
offices where the robot
uses an internal map and
beacons to deliver mail.
//Remote controlled Robot Simulation
//Define Constants
LEFT_ARROW = 37
UP_ARROW = 38
RIGHT_ARROW = 39
DOWN_ARROW = 40
//
MainProgram:
setcolor white,blue
rCommPort 0
rLocate 400,300
Gosub RemoteControl
End
/-Subroutines
RemoteControl:
while true
getkeyE K
if K <> 0
if K = UP_ARROW
if !(rBumper() & 4) then rForward 1
elseif K = DOWN_ARROW
if !(rBumper() & 1) then rForward -1
elseif K = LEFT_ARROW
rTurn -1
elseif K = RIGHT_ARROW
rTurn 1
endif
rectangle 0,0,800,25,blue,blue
xystring 1,2, Bumpers=,rBumper()
endif
wend
Return
//-
two bytes to the target robot. The first
of these two bytes is an ID code to
identify the command and the second
byte is a parameter that is needed for
some commands (the second byte will
be zero if not needed). For example,
the command rForward 10 will cause
the byte 06 to be sent followed by 10,
where 06 is the code for the
forwarding command and 10 is the
parameter representing the distance to
forward. Table 1 shows details of
the data sent by RobotBASIC for each
of the supported statements.
When the robot receives the
two-byte command from the PC, its
hardware-level control program must
execute that command and then send
back exactly five bytes to the PC.
Except in the case of the GPS
command, the first three of these bytes
are the data from the robots bumper
sensors, infrared perimeter sensors, and
line sensors (or zero if some of these
sensors do not exist in the robot).
A robot typically needs to react
quickly to the data from these three
sensors, which is why the protocol
requires it to be returned after each
command. The received information is
stored in a buffer within RobotBASIC
and is provided to the application pro-
gram whenever it uses the functions
rBumper(), rFeel(), or rSense().
The last two bytes (of the five
bytes returned to the PC) are data
that may be required, depending on
the command. For example, rRange()
expects some form of ranging sensor
(sonar, infrared, etc.) to be read and
the value of its reading to be returned
in these last two bytes. The last two
bytes represent a 16-bit integer where
the most significant byte is to be
received first and the least significant
byte second. Data for other sensors
such as the compass or the camera
are also returned in these last two
bytes. The actual devices used by the
robot are immaterial because the
details of the hardware are totally
transparent to the language.
Obviously, the actual hardware-
level control program will be unique for
any given robot, based on its capabili-
ties and configuration. Nevertheless,
we will show a typical shell program
that illustrates the principles involved.
Once you understand this code, you
can write a program catered to
your robots motors and sensors in
a language appropriate for the
microcontroller of your choice.
It can take significant effort
to prepare a custom hardware-level
control program for your robot, but
remember, once this is finished, it
never has to be done again (unless you
add new sensors or change the design
of your physical robot). Hopefully, in
the near future, robot kits will be
available that respond to the Robot
BASIC protocol right out of the box.
Since all of the intelligence and
problem-solving decisions reside in the
RobotBASIC program, you have the
power to create far more intelligent
robots than could ever be done with
a microcontroller-based machine.
With RobotBASIC, you have a virtually
unlimited space for both fixed and
floating-point variables and arrays. You
have modern control structures and
powerful mathematical functions
always available. You can debug all
your code using a simulation and
then use it on a real robot without
translating code or downloading
software. RobotBASIC provides all
this and more, creating a near perfect
environment for developing robot
control programs. We believe this is
the future of hobby robotics.
The Hardware-Level
Control Program
The PBASIC program for the BASIC
Stamp 2 from Parallax (www.Par
allax.com) is an example of an appropri-
ate hardware-level control program. The
program is too long to print here but you
can access it from the SERVO website
(www.servomagazine.com). Most of
the routines in the shell program are not
needed for this example, so only a debug
statement has been inserted to indicate
where the code to carry out necessary
actions should be placed.
For this example, we will assume
that the robot is moved by two servo
motors and that it has four bumper
Although we designed the
RobotBASIC protocol to be easy to
use, we wanted to keep the system
very flexible. It is possible to use our
existing commands in totally different
ways. The rSense( ) function, for
example, is used on the simulator to
read data from the line sensors. If
you were building a firefighting robot
that did not need line sensors but
did need a heat sensor, you could
program your real robot to send back
HEAT data instead of LINE data when
rSense( ) is used. Furthermore, if you
use a red circle in the simulated
environment to represent a fire, then
even the simulated robot could use
rSense( ) to locate and react to the fire.
In order to add even more
flexibility, the standard internal serial
commands support the wireless
communication mode so that you can
create subroutines to handle any
sensor that might come along in
the future or even create your own
personal protocol, if you wish.
ADVANCED USERS
60 SERVO 03.2008
RobotBASIC ID Function Second
Command Code Byte
rLocate 3 Initialize Zero
rForward 6 Move forward Amount to move
7 Move backward Amount to move
rTurn 12 Turn right Degrees to turn
13 Turn left Degrees to turn
rLook 48 Camera (color detect) Right angle (degrees)
49 Camera (color detect) Left angle (degrees)
rRange 192 Ultrasonic ranger Right angle (degrees)
193 Ultrasonic ranger Left angle (degrees)
rCompass 24 Compass Zero
rBeacon 96 Beacon detect Zero
rPen 129 Pen up/down Zero/non-zero
rSpeed 36 Speed control Speed factor
rGPS 66 GPS coordinates Zero
TABLE 1. These commands send a request for action to the target robot.
Blankenship.qxd 2/6/2008 11:01 AM Page 60
switches around its perimeter (the
same as the standard RobotBASIC
simulation). This means that the only
routines we must complete will be
those that control the servo motors
and read the bumper sensors. The I/O
pins used for all the devices will
be specified in the code. We used
Parallaxs EB500 transceiver to provide
the Bluetooth communication for the
robot and a Bluetooth USB adapter
for the PC.
Remember, this is only an example
program to demonstrate the concepts.
You can use any language and any
microcontroller.
Controlling The Real Robot
Once you have installed a proper
hardware-level control program on
your robot, you can use the remote-
control program of Figure 3 (or
any RobotBASIC program) to control
your target machine. This is the
purpose of the rCommPort command
discussed earlier. This command tells
RobotBASIC to use the serial port
specified to communicate with the
microcontroller of the real robot. This
serial port can be a real serial port
or a virtual one that is set up by a
USB device driver or a Bluetooth
device driver.
If the port number specified in the
rCommPort statement is non-zero,
RobotBASIC will immediately start
communicating with the target robot,
according to the protocol described
earlier. As long as the target robot has
the proper hardware-level control
program, the real robot will respond in
the same manner as the simulated
robot. For example, if you are running
the remote-control program and press
the Up-Arrow key, the real robot will
move forward unless the front bumper
is in contact with an object. In the real
mode, the bumper status displayed on
the screen will be that of the real
robot, not the simulation.
Note: If you want to experiment
with the concepts in this article and do
not have and do not wish to buy
wireless transceivers, you will be able
to do so using a direct serial wire
between the PC and the BS2 (or
whatever controller you choose).
This and many other options for
implementing a communication link
between the BS2 and PC are described
in a PDF document that you can
download from www.RobotBASIC.
com. The document will also show
how to experiment with the software
without having to have an actual
robot, just the BS2 and some push
buttons and LEDs. You can also down-
load the programs shown in the article
and the latest version of RobotBASIC.
As you can see, RobotBASIC is a
powerful robot language, but it can also
be used for nearly any control applica-
tion. There are serial and parallel I/O
commands not discussed here that pro-
vide many options for communicating
with external circuitry, thus facilitating
the development of numerous types of
electronic projects. SV
SERVO 03.2008 61
Hitec Robotics invites all humanoid roboticists this April 25th-
27th at EFX (the Electric Flight Expo) at the University of Phoenix
Stadium in sunny Arizona. Humanoid robots of all types and
brands (both homemade and kit-form) are welcome to compete
in the first annual Hitec Robotics Cup.
The Hitec Robotics Cup is composed of three different events:
Robo Speed: Robot sprint of 3m
Robo Mission: An obstacle course with stairs, blocks, and turns
Robo Duel: One on one fighting match between bots
Winners in all three categories will receive cash prizes & trophies!
Full rules and registration details available at our website. If you
have any questions about the event please email
tonyo@hitecrcd.com or go to http://www.hitecrobotics.com
1st Annual Hitec Robotics Cup
62 SERVO 03.2008
When you are creating a
custom robot, there is no manual,
instructions, or tech support. You
start with a blank slate, with
thousands of options available to you.
You have to ask yourself and answer
questions like How big will it be? What
will power it? What materials will I use
for the frame? and many more. With
an unstructured approach, this can
seem like an impossible task. However,
if you approach the problem with a
plan, breaking down the problem into
chunks, it is not as difficult as it seems!
In this series of articles, I will
be discussing how to go through the
entire process of designing and
building a custom robot. Ill show you
how to break the process down into
manageable chunks and see to it that
the idea in your head turns into a reality.
Identify the Goal
Every engineering challenge begins
with a goal or problem statement that
clearly and concisely lays out what you
want to achieve. It should describe what
the purpose and strategy of the robot
is. It should not, however, limit or direct
what the final design will be.
My background is in combat
robotics, so I will be discussing the design
process for that type of robot. As an
example, I have decided I want to build a
competitive, offensive based robot that
has the capability to knock out other
robots by causing large amounts of
damage. Notice that I didnt specify how
I wanted to cause large amounts of dam-
age; there should be no how in a prob-
lem statement only a what or why.
The most common example of a
good problem statement is for mowing
the lawn. We need a better product to
mow the lawn is flawed, because it lim-
its you to only ways to mow the lawn.
However, We need a new, innovate
way to keep lawns looking good opens
up a completely new set of possibilities.
A good goal statement will help to
direct your design from start to finish.
Design Specifications
After identifying qualitatively what I
want the robot to do, I need to decide on
a set of design specifications. This is a list
of criteria that describes the robot quanti-
tatively and will begin to answer questions
like How many wheels? What is the
maximum weight? How fast will it move?
It needs to be as specific as possible to
make the design process easier later on.
The best type of specifications are
ones that can be measured, so that
throughout the design process you can
evaluate how well your design is meet-
ing your goal. For my planned robot,
the design specifications are as follows:
Combat robot for fighting in national
competitions
Middleweight weight class (120 pound
maximum weight)
Four-wheel drive
Designing and Building a
ROBOT from
SCRATCH
D
esigning and
building a
custom robot
for the first time can be
a very daunting task.
Robots consist of many
subsystems, all of which
must work seamlessly
in order for the robot to
function properly. The
challenge in designing
and building a robot
is not only in fully
understanding these
subsystems and making
them work together, but
doing it in a way that has
never been done before.
b y B r i a n B e n s o n
Benson.qxd 2/5/2008 6:12 PM Page 62
Inboard wheels
Tank steering
Minimum drivetrain speed of 6 mph
Minimum 1/4 ground clearance
Five minute run time
As small a size as possible in all
dimensions
Importance of subsystems
40% weapon
15% armor
20% drive train
25% reliability
Comply with all of the Robot Fighting
League and BotsIQ rules
These design specifications were
chosen for a variety of reasons. I decid-
ed on the 120 lb middleweight class for
personal reasons. If you are building for
a competition, you will probably have a
maximum weight. If it is just a hobby
robot, then weight may not be as
critical. The number of wheel drive that
you choose will directly affect how your
robot handles. A two-wheel drive robot
will turn quicker and more efficiently,
but will not drive very straight. It will be
what many builders call squirrelly.
While a four-wheel drive robot will be
much easier to control, it will be heavier
and take up more space. I decided on
four-wheel drive because all of my
previous robots have been two-wheel
drive and very difficult to control (which
I wanted to avoid). I decided on the five
minute run time because that is the
longest match time in competition.
Battery life is very important in all
mobile robots, so be sure to specify how
long you expect to use it on one charge.
The purpose of specifying the
importance of the different subsystems
was to help determine how to distrib-
ute the weight, and where to make
sacrifices in the design. Engineering is
all about study and compromise.
Real life restrictions prevent you
from having everything ideal, so
you have to study all of the
available options and compro-
mise on some so that others are
optimized. The common saying is
Good, Fast, Cheap. Pick two.
Idea Development
The next step is to brainstorm on
how these design specifications and
goals will actualize. There are many
different techniques for brainstorming.
These include random idea generation,
researching previous work for inspiration,
combining previous ideas, and many
more. Overall, you want to come up with
as many ideas as possible, without reject-
ing any. In the case of my new robot, I
want to cause high amounts of damage
capable of knocking the opponent out.
So, I came up with the following list:
1) Vertical bar or disk spinner
2) Horizontal bar or disk spinner
3) Hammer
4) Full-body spinner
5) Powerful flipper
6) Clamp and drill
7) Crusher
8) Thawck
9) Choosing an idea
There are many different ways to
choose an idea, and they range from
simply choosing one to analytical
methods. While you may have already
decided which idea is best, it is a good
idea to compare them systematically,
anyway. There are many ways to do
this, but I will outline my favorite way.
The first step involves rating each
design in a number of categories. First,
you must choose a set of categories to
rate each idea by. These will probably be
closely related to your problem statement
and design criteria. In this case, I want
maximum damage and ability to knock
out other people, competitiveness, reliabil-
ity, and I want it to be feasible given real
world limitations. So, I will create a table
like the one in Figure 1 in which I rate how
well each design satisfies each require-
ment. I will use a rating system from zero
to three; three being the highest.
One option is to assign a weight to
each category by importance. In this case,
I will give Damage and KO Ability a
weight of 200%, and the rest a weighting
of 100%. This means I think that
category is twice as important. Then, the
ratings can be totaled to determine how
they compare. Remember, each situation
will have its own categories, so choose
ones that you feel are most important.
The second step involves comparing
two ideas at a time. This part is more sub-
jective in that instead of having different
categories, you simply do a comparative
rating in what is called the pairwise
comparison chart. You create a table
with the first row and first column having
each of the ideas. You then go through
the chart, comparing two ideas (the one
on the left relative to the one above) at a
time numerically at their intersection as
shown in Figure 2. In this case, I will
assign a zero if the idea above is better.
If the idea to the left is better, then I will
assign a one. While deciding which one
is better, I want to keep my problem
statement and design criteria in mind.
The rating chart setup showed
that both the vertical and horizontal
spinners would work best. This made
sense and agreed with my initial incli-
nation towards a spinner. The pairwise
comparison reinforced that a spinner
would be best, and showed that a hor-
izontal spinner is better then a vertical.
Not only did this process agree with
my initial inclination, it analytically
proved that it was correct. However, this
doesnt mean that it was pointless.
There are times and they have hap-
pened to me when this process has
convinced me to choose another design.
SERVO 03.2008 63
Damage and Competitiveness Reliability Feasibility Total Sum
KO Ability (200%) (100%) (100%) (100%)
Vertical 2 3 3 3 13
Spinner
Horizontal 3 3 1 3 13
Spinner
Thwack 0 2 3 3 8
Crusher 2 3 2 1 10
Flipper 1 2 2 2 8
Hammer 1 2 2 2 8
Clamp and Drill 2 1 2 0 7
Rating Chart
FIGURE 1. Each idea is rated
in different categories which
can be weighted, from 0 to 3,
3 being the best. This allows
you to compare the ideas in
terms of specific attributes.
In this case, I now have to choose
what kind of horizontal spinner I want:
bar, disk, chain, etc. You also may find
yourself in a situation where now you
have to narrow your design down further.
Simply apply the same procedure again.
I researched how previous robots
with different types of horizontal
spinners had performed, noticing their
strengths and weaknesses. I then used
this information for a rating chart and
pairwise comparison, and decided on a
horizontal overhead spinning bar,
much like an upside-down lawnmower.
This allows the weapon to be effective
around the entire robot, therefore also
acting as a type of armor.
In Review
Starting with only the thought I
want to build a robot, we came up
with a goal for the robot and a specific
set of criteria we wanted it to meet.
Next, we brainstormed many different
ideas, compared them analytically, and
chose the one that was best. This
process allowed us to create a paper
trail of how we got to the final idea,
which can be important down the road.
With a specific idea in mind, the
next step is to begin the actual design.
In Part 2 of this series, I will be
discussing how to determine what
you need for parts, how to choose
them, developing your design, and
more. For more information on myself
or my robots, visit www.robotic-hob
bies.com. SV
64 SERVO 03.2008
Vertical Horizontal Thwack Crusher Flipper Hammer Clamp and Total Sum
Spinner Spinner Drill
Vertical x 0 1 1 1 1 1 5
Spinner
Horizontal 1 x 1 1 1 1 1 6
Spinner
Thwack 0 0 x 0 0 0 0 0
Crusher 0 0 1 x 1 1 1 4
Flipper 0 0 1 0 x 1 1 3
Hammer 0 0 1 0 0 x 1 2
Clamp and Drill 0 0 1 0 0 0 x 1
Pairwise Comparison Chart
FIGURE 2. Using a pairwise comparison
chart allows you to compare two ideas
at a time in order to rank order all of
them.
Advanced Course 3 Days $1995
Program and Control:
GPS Receivers
RS232, 485 Serial Communications
Wireless Communications
IEEE Zig Bee, Bluetooth
Distance Sensors
RFID Readers and Tags
Motor Control W/Feedback
Minneapolis, MN, Jul. 21-23
Los Angeles, CA, Aug. 18-20
Newark, NJ, Sep. 22-24
San Diego, CA, Oct. 6-8
Check the web for more dates
Intro Course 3 Days $1995
Program and Control:
LED and LCD Displays
DC, Servo, and Stepper Motors
Temperature and Light sensors
RS232 Serial Communications
ADCs and EEPROMs
Timers, Interrupts, & Clocks
Visual Basic Interfaces
Columbus, OH, Jan. 21-23
Los Angeles, CA, Feb. 4-6
Huntsville, AL, Feb. 18-20
Newark, NJ, Mar. 3-5
Check the web for more dates
WoW!
Finally,
Hands-on
Training for
PIC
Microcontrollers
and PIC BASICPRO
!
RCG
Research
Innovative Ideas in Electronics Design
www.RCGResearch.com
Phone (800) 442-8272
PIC and PIC BASICPRO are trademarks or registered trademarks of Microchip Technologies, Inc.
RCG Research, Inc. is not affiliated with Microchip Technologies, Inc.
ROBOTS ARE ART
SERVO 03.2008 65
The Floatation Center and Art
Gallery in northern CA held a
robotics art show recently.
Read this months Appetizer
on page 78 for details.
Menagerie Mar08.qxd 2/6/2008 3:58 PM Page 65
66 SERVO 03.2008
Electronic Parts & Supplies
Since 1967
www.c-stamp.com
L
O
W

C
O
S
T
!
CONTROLLERS & ROBOT KITS SENSORS DISPLAYS
Free
Expert
Technical
Support
1 (800) 985-AWIT
E
a
s
y

t
o

C
o
n
n
e
c
t
L
O
W

C
O
S
T
!
LOW COST
Professional SW Dev. Tools
For the finest in robots,
parts, and services, go to
www.servomagazine.com
and click on Robo-Links.
Study online, at your own pace and earn your certificate as a...
www.online-tech-training.com/nv
Programmable Logic Controllers
Technician
RobolinksMAR08.qxd 2/6/2008 1:29 PM Page 66
I
n my experience, definitions of life
tend to raise more questions than
they answer. Countless descriptions
exist from different perspectives and
for different purposes. For example,
the metabolic definition states that liv-
ing things have the ability to take ener-
gy from their environment and change
it from one form to another. The genet-
ic definition focuses on evolution as the
distinctive property of living systems.
The biochemical definition views hered-
itary information storage in DNA as the
key. Nobel prize-winning physicist
Erwin Schrdinger arrived at what has
been called the negentropic defini-
tion of life, which observes that living
things export entropy in order to keep
their own entropy at a low level.
Biologists Humberto Maturana
and Francisco Varela coined the
term autopoiesis to describe life as
an autonomous, self-organizing,
and self-producing system.
Similarly, theoretical biologist Stuart
Kauffman defines a living thing as
an autonomous agent, something
that can both reproduce itself and
do at least one thermodynamic
work cycle.
Unfortunately for the semanti-
cians, no definition seems to be
satisfactory to everyone. Viruses, fire,
artificial life, and of course the
possibility of yet-unknown extraterres-
trial life forms, all pose challenges to
potential all-inclusive definitions.
Though some definitions of life
may accommodate the inclusion of
such bizarre organisms as software
algorithms, this idea is gen-
erally not greeted with pop-
ular accord. In this sense,
artificial life questions the
very meaning of life itself,
forcing us to examine the
basic semantic distinction
between the living and the
non-living. The confusion brings us
back to the fundamental philosophical
debate between Vitalism and
Materialism: Is there, in fact, a vital
spark which ignites the living, meta-
physically separating it off from the
non-living? Or is life simply a property
of a very complex emergent system?
In my opinion, the confusion
unfolds as a question of medium speci-
by Heather Dewey-Hagborg
Have you ever questioned the meaning of life? Im not talking about existentialism
here, but the meaning of the word life what it means to be alive? Have you ever
wondered, for example, if there was some way, some possibility, that an electronic
creation (your electronic creation) could one day be deemed alive?
DIFFERENT
BITS
DIFFERENT
BITS
ARTIFICIAL LIFE:
PART 1 INTRODUCTION TO GENETIC ALGORITHMS
FIGURE 1. Image from Ken Rinaldos
interactive robotic installation
Autopoiesis. In his own words:
Autopoiesis is self making, a
characteristic of all living systems
which was defined and refined by
Francisco Varela and Humberto
Maturana. The interactivity engages
the viewer/participant who, in turn,
effects the systems evolution and
emergence. This creates a system
evolution, as well as an overall
group sculptural aesthetic. (Used
with permission of the artist).
SERVO 03.2008 67
DifferentBits.qxd 2/6/2008 10:50 AM Page 67
DIFFERENT BITS
68 SERVO 03.2008
ficity and it runs throughout the field
of artificial intelligence studies. Can a
thing be said to be alive, intelligent, or
conscious, if it is made of different
stuff than we expect? Marvin Minskys
classic dictum that a brain is simply a
machine made of meat allows us to
imagine that minds can be made in any
medium, and whether the substance is
meat or silicon is irrelevant to the out-
come: the behavior of the machine.
Artificial life is the study of
artificial systems that exhibit behavior
characteristic of natural living systems.
It is the quest to explain life in any of
its possible manifestations, without
restriction to the particular examples
that have evolved on earth. [1]
As Minsky invited us to imagine
the possibilities of computer intelli-
gence, Langton invites us to ponder
the possibilities of silicon life forms.
In this next series of Different Bits
columns, we will take up and explore
these ideas which blur the traditional
boundaries of biology and engineering
and call conventional definitions of
life into question. We will venture
into the world of artificial life: genetic
algorithms, cellular automata, and
emergent systems.
Genetic Algorithms:
An Introduction
The first part of this series will
focus on genetic algorithms. This
column will introduce the key concepts
behind the technique. Next time, we
will look at a hello world example,
and a third installment will cover the
topic of interactive evolution.
Although genetic algorithms are a
software method, we will implement
our applications using microcontrollers
and physical computing interfaces. As
with all columns in this series, a
basic knowledge of electronics and
programming is assumed. But, before
we jump into algorithms, we need a
little background on the biology that
they are derived from.
Evolution
Evolution is most simply defined
as an accumulation of changes in
a species over generations, which
eventually leads to branching of the old
species into multiple new species. The
theory proposes that there was at one
time an original species which has
transformed and split over time into
the diversity of life we know on the
planet today.
Adaptation is the process which
allows organisms to inherit useful traits
from their ancestors. It is the ability of
a species to change over time in a feed-
back loop with its environment. There
are three main components to adapta-
tion: heredity, selection, and variation.
Heredity
DNA is a molecule which carries
instructions to inform cell develop-
ment. It is biologys form of long term
information storage for the blueprints
of living organisms. Genes are specific
regions of DNA that code for physical
traits. Heredity is the ability of an
organism to pass on these genes and
therefore traits from generation to
generation through reproduction.
Every living organism has a genotype,
which is the complete set of all
its genes. In correlation with the
genotype, an organisms phenotype is
the physical expression of its genes as
they interact with the environment.
Selection
The process of evolution utilizes
three primary selection methods:
natural selection, genetic drift, and
gene flow. Natural selection is the
likelihood of an organism with a
beneficial mutation in its genome to
survive and to reproduce. The concept
of natural selection rests upon the
observation that all organisms inhabit
an environment of limited resources
and must, on some level, compete with
each other to survive. This is the idea
behind the oft-quoted phrase survival
of the fittest, though some have
suggested the phrase should be revised
to survival of the reproducers since it
is irrelevant to evolution if an organism
survives without reproducing; without
passing on its genes. [2]
Genetic drift is the fluctuation of
inherited traits in a population of organ-
isms due to chance events. For exam-
ple, one individual in the population
reproduces significantly more than its
neighbors, with no relation to fitness.
Gene flow is the migration of
organisms from one region to another
coupled with reproduction in the new
environment. Gene flow allows genes
to travel from one population to
another across great distances.
In this article, we will be focusing
only on natural selection, as this is the
inspiration for the genetic algorithm.
Variation
Variation consists of mutation and
recombination. Mutation is a little bit
of randomness thrown into the mix. It
is the small possibility that during
reproduction, the new offsprings DNA
will mutate to form a new gene and
therefore, possibly, a new trait.
Mutation is caused by environmental
factors such as radiation and also
chance factors such as errors in DNA
replication during reproduction.
Recombination is the method by
which sexually reproducing organisms
(like humans) produce new offspring
which contain random recombinations
of their parents genes. Recombination
is the key to evolutions success as it
allows children to be different than
their parents. This means beneficial
genes tend to propagate through time
whereas harmful genes tend to be
evolved out.
FIGURE 2. A single gene, highlighted in a strand of
DNA, extracted from its home in a chromosome
(modified from Wikipedia).
In contrast, asexually reproducing
organisms such as bacteria reproduce
through cloning: a process which
excludes recombination and therefore
relies on selection and mutation alone
to adapt. This leaves the population
with less genetic diversity, a slower
rate of adaptation, and therefore
greater vulnerability during sudden
environmental shifts.
Algorithms
With nature as our inspiration, we
can begin to abstract the processes of
evolution, and to evolve populations of
algorithms to solve difficult problems
computationally. Beginning with a
specific problem and goal, a genetic
algorithm will evolve a population of
candidate solutions which compete on
the basis of their fitness, or proximity
to the goal. The algorithm is iterated
such that each time through a loop the
best candidate solutions breed to form
the next generation.
Utilizing operations inspired by
biological cross-over and mutation, evo-
lutionary algorithms are capable of
effectively finding solutions to problems
with search spaces which are too large
to tackle using a brute force search.
Like biological evolution, they do not
guarantee to result in the best solution
to the problem; rather, they promise a
solution which is good enough; which
meets the minimum criteria for success.
The genetic algorithm was formal-
ized by John Holland in his 1975 book
Adaptation in Natural and Artificial
Systems. The algorithm is simple:
1) Create an initial population of
candidate solutions.
2) Evaluate the initial population.
3) While no member of the population
meets the criteria for success (breed-
ing loop):
a) Select individuals into a mating
pool.
b) Create a new population using
crossover and mutation.
c) Evaluate the new population.
The most difficult part of working
with genetic algorithms is generally
finding a good mapping between the
genotype and phenotype of the popu-
lation; in other words, figuring out how
to represent the problem as a string of
numbers. Genetic algorithms historical-
ly use a binary representation of the
possible solution to form a kind of
computational DNA. An example geno-
type might look something like this:
1100 0010 1111 0001
The DNA representation must be
able to adequately describe a solution
and be modular enough to progress
using crossover and mutation.
For example, if you want to evolve
organisms with specific hair and eye
colors, you would represent it with two
binary numbers big enough to hold the
number of possible options. If possible
hair colors are brown (0), red (1),
blonde (2), and black (3), the gene
representation size would be 3 or
binary 11. Another gene could hold
four possible values of eye colors:
brown (0), blue (1), green (2), and
hazel (3). So, our DNA encoding would
look something like this:
1) Population
To create the initial population of
candidate solutions, first we decide how
large the population should be. There is
often a trade-off between population
size and iterations; the more population
members you have, the more time it
takes to compute a generation of evolu-
tion. But if you have too few members
of your population, you may not
have enough diversity to find a good
solution. For our example, we will start
with a population size of 10, so we
create 10 arrays of four random bits:
1. 1001
2. 1111
3. 1001
4. 0000
5. 1010
6. 1110
7. 1000
8. 0011
9. 0110
10.1101
2) Evaluation
The next step is to determine a
criterion for success. In this example, I
am randomly deciding that the criteria
will be having red hair and hazel eyes
the genetic sequence 0111. Now, we
walk through each member of the
candidate population and measure
how close they are to the solution we
are trying to evolve. For example, the
first member of the population has the
genome 1001 or 2, 1 which translates
into blonde hair and blue eyes. Neither
of these traits correspond to the ones
we are attempting to evolve, so this
candidate gets a score of zero out of a
possible two points.
When we examine the second
member of the population in the
same manner, we see that they
have black hair and hazel eyes, giving
them a score of one. After evaluating
each member of the population like
this, we end up with the following list
of scores:
1. 0
2. 1
3. 0
4. 0
5. 0
6. 0
7. 0
8. 1
9. 1
10.0
No member of the population
meets our criteria for success, so we
enter the breeding loop.
SERVO 03.2008 69
DIFFERENT BITS
FIGURE 3. DNA representation of the traits. This
example shows how four bits can code for four
different hair colors and four different eye colors.
3) Breeding
a) Selection
The first step in breeding is
selection; we need to determine which
members of the population will survive
to reproduce, and how many offspring
each member will have. This is the
correlate of natural selection in biology
and so we need a method
which allows the most suc-
cessful genes to continue on,
while still ensuring genetic
diversity. The individuals who
survive to reproduce comprise
the mating pool. Their off-
spring will become the new
population and will complete-
ly replace the previous generation.
The simplest method of selection is
called truncation. This technique
determines in threshold of selection
and only allows individuals who meet
the threshold to mate. In our example,
our threshold would be a fitness score
of 1, allowing three members of the
population to breed. The problem with
this is that in many situations (like
our example), it will lead to a very
limited and insular gene pool. By
only allowing the fittest to breed,
we may never have enough genet-
ic diversity to arrive at a solution.
Proportional selection solves
this problem by allowing candidates a
probability of mating which is propor-
tional to their fitness. It works by copy-
ing each member of the population into
a mating pool a number of times
equivalent to their rating. There are
more complicated equations for doing
this, but microcontrollers often have lim-
ited math and working intuitively often
works just as well. In the above exam-
ple, DNA with a score of zero would be
copied into the gene pool once. DNA
with a score of one are doing signifi-
cantly better; they are 50% correct, so
we want a lot more copies of them. You
can adjust the parameters for how
many copies to add, but I would start
with five, or 50% of the population size.
70 SERVO 03.2008
DIFFERENT BITS
FIGURE 4. Example of the crossover
operation during the breeding cycle. An
index is randomly chosen as a splitting point
in both parents DNA. Opposite segments
are then swapped to create two new DNA
sequences of the same length. The new
sequences then become child 1 and child 2.
b) Crossover and mutation
Now that we have a mating pool,
we are ready to reproduce. The most
common method is to randomly
choose two individuals at a time from
the mating pool. The crossover method
is inspired by the biological process
of meiosis. It consists of randomly
choosing one or more splicing points in
both parents DNA and swapping
alternate pieces. The process results in
two offspring which are a combination
of both parents DNA.
The children are then subject to a
probability of mutation which random-
ly flips a specified percentage of bits. It
is important not to mutate every
offspring, so the new generation is
generally subject to a specified proba-
bility of mutation, often starting
around 1%. In the example above,
with a population size of 10 six bit
DNA strands, only an average of two
bits for every five generations would be
subject to mutation. Mutation provides
that extra bit of randomness and
genetic diversity which proves often
crucial for getting results.
c) Evaluate
Once enough children are pro-
duced to entirely replace the previous
generation (in our example, 10), their
fitness is evaluated in exactly the
same manner as their parents. Each
child gets a fitness score, and unless a
solution is found, the children
reproduce based on their fitness to
form a new generation ... which is
evaluated, and so on for eternity, or
until the exit condition is met.
Conclusion
Genetic algorithms can be used to
solve complex problems, to emulate
biological processes, or to create com-
plex systems of autonomous agents. In
the next installment of Different Bits,
we will implement a simple example of
the genetic algorithm to illustrate just
how useful and fun they really are. SV
Heather Dewey-Hagborg can be
contacted via email at heather.
servomagazine@gmail.com
CONTACT THE AUTHOR
DIFFERENT BITS
SERVO 03.2008 71
[1] Artificial Life I (Christopher Langton)
[2] The Computational Beauty of
Nature (Gary William Flake)
Artificial Intelligence: Structures and
Strategies for Complex Problem
Solving (George F. Luger)
Metacreation: Art and Artificial Life
(Mitchell Whitelaw)
The Tree of Knowledge (Humberto
Maturana and Francisco Varela)
Introduction to Artificial Life
(Christoph Adami)
REFERENCES
72 SERVO 03.2008
I
vividly remember the excitement of
building my first electronics kit. Okay,
so the kit was an old surplus reject that
used vacuum tubes instead of transistors.
But to an 11-year-old interested in the sci-
ence of electronics, all that mattered was
the scores of tiny parts packed into indi-
vidual numbered plastic baggies, and the
smell of smoldering solder as I methodi-
cally attached one wire to the next.
The thing didnt work when I first
turned it on in fact, sparks flew and
it tripped the circuit breaker when I
flicked the switch. But all was well after
my dad looked over my handiwork,
re-soldered some cold joints, and
corrected the connections I had wrong.
For years after, I enjoyed listening to
my phonograph player through the
three-tube amplifier that I built.
Kits are ideal for those wanting to
build things but that arent too keen on
having to round up all the parts before-
hand. Just like with electronics, robot kits
help you make your own bot, saving you
lots of time and even money. Unless
you have parts left over from a previous
project, its usually cheaper to build a
robot from a pre-packaged kit, because
you dont have to buy all the bits and
pieces from several different sources, sav-
ing your gasoline bill and shipping costs.
In this installment of Robotics
Resources, well cover a selection of
robotics kits and kit sources. There are
dozens and dozens of robot kits, and
hundreds of online retailers selling them,
so theres no way we can cover them all.
But the assortment described here
should give you a good idea of whats
available, and where you can get it.
All-in-One or Base Kits
Not all robotic kits are wholly
self-contained, providing absolutely
everything you need to build your first (or
second or third or ...) robot. Some are
all-inclusive, such as the Parallax BOE-Bot,
which includes the hardware, software,
and computer programming essentials to
construct a functioning autonomous
vehicle. These kits serve as an excellent
introduction to robotics, and are particu-
larly well suited for those just starting out.
Then there are the base kits, such as
those from Zagros Robotics or my own
small manufacturing company, Budget
Robotics. These provide the hardware
the mechanicals including chassis, frame,
wheels, and usually motors but not
the electronics. You are expected to add
those yourself. This approach to robotic
kits allows you to reuse your favorite
microcontroller or other electronics, or
mix-and-match components so you get
exactly what you need. These kinds of
kits generally require a little more experi-
ence, but offer the flexibility of selecting
just the components you want to use.
Too, there are variations within the
theme of the all-in-one kit and the base
kit. Modular kits such as the SuperDroid
Robots Trekker line offer a means to
customize elements, including drive
components and sensors. The modular
components let you add custom features
if and when you want them a sort of
attachments system for desktop robotics.
Construction set kits, which include
the Lynxmotion Servo Erector Set,
the Robotis Bioloid, LEGO Mindstorms,
and the Vex from Vex Robotics, are
comprised of multiple interchangeable
building parts. The parts are designed
around a common set of motors and
other electronics. With a different
arrangement of components, you can
build two-, four-, or six-legged walking
robots, robots that run on two or more
wheels, or tracked tank-like robots.
On the other end of the spectrum
are base components, which you then
use to build a robot of your own design.
The base components are pre-cut and
pre-drilled, saving you the labor of
shaping these parts in your garage. For
example, if youre building a small desk-
top robot, you can find pre-cut discs and
other base parts made of metal, plastic,
and even wood. Added to the base
might be motor mounts and stand-offs to
construct a multi-tiered chassis. You are
expected to supply the motors, wheels,
battery and holder, microcontroller, and
other parts to complete the system.
Across the spectrum of kit types,
nearly all allow some form of modifica-
tion and customization. Modular designs
work best with components made for
them, and youll get the most profession-
al-looking results if you use components
specifically designed for the robot. Even
so, in nearly all kits you can fashion your
own solutions, and attach them to your
creation using small nuts, bolts, hook
and loop fasteners or sticky-back tape.
Bases made of wood or plastic are easier
to rework using ordinary shop tools.
Rolling, Crawling,
Walking Bots
The three most common robot
Robot Kits for
Easier Robotics
Tune in each month for a heads-up on
where to get all of your robotics
resources for the best prices!
RoboResources.qxd 2/6/2008 10:48 AM Page 72
designs wheeled, legged, and tracked
are all well represented in the available
kits. Robots with wheels are among the
least expensive, because the mechanics
of attaching a wheel to a motor is fairly
simple. Most robot kits are the familiar
two-wheel design, where a third
(sometimes fourth) wheel or mechanical
skid serve for balance. For the smaller
and lighter desktop robot class (about
10 inches or under), two drive wheels
are directly connected to their motors,
which are attached to the robot chassis
using hardware or other fastener system.
The robot is steered by controlling the
speed and direction of each wheel.
Tank treads are also a popular robot
locomotion choice. To save cost, most
kits use all-rubber treads (often from the
Tamiya Track & Wheel Set) available from
hobby stores. The set includes multiple
rubber segments that can be connected
together to make different lengths of
tread, as well as an assortment of plastic
roller and drive sprocket parts.
More advanced (and expensive)
treaded kits use specially made metal or
plastic parts to form the tracks. Two exam-
ples are the track kits from Lynxmotion
and Vex Robotics. Both use interlocking
plastic pieces, allowing you to construct a
track of exactly the length you want.
Legged robots are among the most
challenging and rewarding designs, and
they tend to also be the most expensive
as kits. Kits may be made of plastic,
metal, or wood, such as the Talrik line
from Mr. Robot (also available in ABS
plastic). Nearly all of the metal kits used
stamped parts, where the metal is cut
out and bent using specialty tools. This
saves the cost of custom-machining
metal parts. A notable exception to this
rule is the Penguin line of two-legged
walking robots from Parallax.
Finding Just the Right Kit
No one robot kit is ideal for
everyone. Picking the perfect kit involves
deciding what you want your robot to
do, and balancing that with your experi-
ence level and budget. Many of the kits
available from the sources below are in
the $50 to $250 range, but there are oth-
ers that cost upwards of $3,000 or more.
The difference is largely the capability of
the robot and the parts that are included.
Obviously, you will pay more for a kit that
is all-inclusive and is bundled with a
microcontroller, sensors, and other parts.
Let others guide you in picking a
kit thats perfect for you. If youre at
school, ask friends or your teacher for
recommendations. On the Internet,
check out newsgroups and forums,
and ask for opinions. If youre lucky
enough to have a robot users group
near you, attend a meeting and seek
the expert advice of other members.
Be sure to be honest about your skill
level and budget; you dont want to be
talked into getting a kit thats over your
head or your pocket book!
Sources
Acroname, Inc.
www.acroname.com
Acroname carries numerous kits
and parts. Also provides a wide selec-
tion of sensors and microcontrollers.
Active Robots
www.active-robots.com
Selection of robots and robotic
construction products, including
microcontroller boards designed with
amateur robots in mind.
Arrick Robotics
www.robotics.com
Though mainly a designer and
builder of automation systems, Arrick
Robotics also sells a small robot
platform, the ARobot. The platform
includes a microcontroller board based
on the BASIC Stamp from Parallax.
A-Wit
www.c-stamp.com
Offers different versions of
the BOL-BOT kit plus accompanying
accessories.
Battle Kits
www.battlekits.com
Heavy-duty designs for those into
competition (battle) robotics.
Budget Robotics
www.budgetrobotics.com
My own company, Budget Robotics
sells a narrow range of products, with
an emphasis on structural elements,
such as robot bodies and frames.
CrustCrawler
www.crustcrawler.com
US-based reseller of the Bioloid
kits, as well as their own custom-made
metal robot kits. Extensive site with
plenty of how-to, programming, and
background information on using their
custom designed robotic kits.
Electronics123
www.electronics123.com
Offers various kits for "future
engineers" including a robot car, frog,
ladybug, solar racer, and many more.
Hitec Robotics
www.robonova.com
Makers of the ROBONOVA-1. Also
offer controllers, servos, sensors, and
other accessories.
Hobby Engineering
www.hobbyengineering.com
Full assortment of controllers,
sensors, and other components for
robotics, as well as bases, starter kits,
and construction components.
HVW Tech
www.hvwtech.com
Canadian Hobby electronics
retailer that manufactures unique Sharp
distance sensors and resells Parallax, PIC,
Atmel, and PICAxe products.
Images Co.
www.imagesco.com
Product line includes an assort-
ment of unusual robotic bases, such as
a golf swing robot, a pan mechanism,
six-legged walking robot made of clear
plastic, and small gripper.
Jameco/Robot Store
www.robotstore.com
Reseller of robot wares, plus
custom kits of their own creation,
including the OctoBot Survivor and
Muscle Wire (shape memory alloy)
articulated armatures.
Joinmax/MCIIRobot
www.mciirobot.com
US distributor of Joinmax kits;
SERVO 03.2008 73
RoboResources.qxd 2/6/2008 3:25 PM Page 73
74 SERVO 03.2008
includes a robot arm and grip kit with
six degrees of freedom.
Junun.org
www.junun.org/MarkIII/
Resellers of the Mark III mini-sumo
robot kit, originally designed by the
Portland Area Robotics Society (PARTS)
for mini-sumo competitions. The kit
includes all hardware, motors, wheels,
and electronics. The website also sells
low-cost sensors, such as the Sharp IR
proximity modules, and various support
electronics common in amateur robotics.
Kadtronix
www.kadtronix.com
Starter robot bases designed for
expansion; products include the
Workman Mobile Robot Platform and a
variety of metal base frames.
LEGO Mindstorms
mindstorms.lego.com
Informational page for the LEGO
Mindstorms sets. Be sure to check out
the user-to-user forums to see what
other LEGO builders are up to.
Lynxmotion
www.lynxmotion.com
Lynxmotion offers complete robot
kits, as well as a unique Servo Erector
Set, a collection of brackets, and other
parts for building custom robots using
standard-size R/C servos.
Machine Science
www.machinescience.com
Offers an expandable metal robot
base which includes motors, wheels,
microcontroller, and other electronics.
Maximum Robotics
www.maximumrobotics.com
Kits for building small robots, plus
controller boards and development
tools for microcontroller-based robotics.
Milford Instruments Ltd.
www.milinst.com
Based in the UK but ships worldwide,
Milford offers the StampBug walking
robot, Big Foot two-legged walking robot,
and Alex, an animatronic/robotic head.
Mr. Robot
www.mrrobot.com
Products include: robot kits (wood
or plastic), servos, DC gearhead
motors, infrared and ultrasonic
sensors, wireless mini color camera,
Fischertechnik construction kits, and
omni-directional wheels.
Norland Research
www.smallrobot.com
Kit of parts for turning a Texas
Instruments scientific calculator into a
robot! Also sells S.A.M., a fun plastic
add-on base for an existing robot.
Parallax
www.parallax.com
In addition to microcontrollers such
as the BASIC Stamp, Parallax offers
numerous robot kits, including the BOE-
Bot (wheeled) and Penguin (walking).
Also offers robot components, such as
wheels, servos, and sensors.
Pololu
www.pololu.com
Makers of small plastic robot bases
designed for use with Tamiya motors.
Also sells servos, wheels, and treads
and sprockets for building tracked
robots, as well as sensors, microcon-
trollers, and other electronics.
qfix
qfix-shop.de
German-based manufacturer and
seller of small robot kits for education.
Website is in Deutsch and English.
RB Robotics
www.rbrobotics.com
New owners of the RB5X education-
al robot. As noted on the website, The
RB5X robot has been produced for over
20 years. With a proven track record and
a developed set of teaching aids, it is a
great choice for any classroom. Full kits
and individual parts available.
Robix
www.robix.com
Aluminum robot kits, primarily
designed for educational use. The
aluminum pieces and servos can be
constructed in numerous ways to make
a variety of robotic forms.
Robodyssey
www.robodyssey.com
Robot kits including both wheeled
and walking designs.
Robotis
www.robotis.com
Manufacturers site for the Robotis
line of robot construction sets and
parts. Check out the dealer pages to
find a reseller near you.
Robotshop
www.robotshop.ca
Resellers of Fischertechnik and
other robot construction sets.
Robot Kits Direct
robotikitsdirect.com
Plastic robot kits in various skill
levels, from beginner to advanced. An
example is the OWI-991K Weasel,
which combined touch sensing and a
basic wall hugging behavior.
Solarbotics
www.solarbotics.com
Primary retailer of BEAM robots
a simple is better approach to
design. Products include various light-
attracted bugs and walking robot kits,
motors, solar cells, and electronics.
Total Robots Ltd.
www.totalrobots.com
UK-based seller of various robot
kits, including BEAM, Lynxmotion, and
general robotics components (motors,
GPS modules, microcontrollers, wheels,
and much more).
VexLabs Robotics Kits
www.vexlabs.com
Makers of the VEX Robotics system.
Zagros Robotics
www.zagrosrobotics.com
Robot bases, large (their designs)
or small (resold from Pololu). An
example base is the Max 14; 14
round ABS plastic with heavy duty
gearmotors and wheels. SV
Gordon McComb can be reached via
email at robots@robotoid.com
CONTACT THE AUTHOR
RoboResources.qxd 2/6/2008 3:26 PM Page 74
Robot Builders Cookbook
by Owen Bishop
This is a book for
first-time robot
builders, advanced
builders wanting to
know more about
programming robots,
and students in
further and higher
education tackling
microcontroller-
based practical
work. They will all
find this book a unique and exciting source
of projects, ideas, and techniques to be
combined into a wide range of fascinating
robots. $29.95
123 Robotics Experiments
for the Evil Genius
by Myke Predko
If you enjoy tinkering
in your workshop and
have a fascination for
robotics, youll have
hours of fun working
through the 123
experiments found in
this innovative project
book. More than just
an enjoyable way to
spend time, these
exciting experiments also provide a solid
grounding in robotics, electronics, and
programming. Each experiment builds on
the skills acquired in those before it so
you develop a hands-on, nuts-and-bolts
understanding of robotics from the
ground up. $24.95
FIRST Robots: Aim High
by Vince Wilczynski / Stephanie
Slezycki
This book looks at
30 different robot
designs all based on
the same chassis,
and provides in-
depth information
on the inspiration
and the technology
that went into build-
ing each of them.
Each robot is fea-
tured in 6-8 pages providing readers with a
solid understanding of how the robot was
conceived and built. There are sketches,
interim drawings, and process shots for each
robot. $39.95
Mechanisms and Mechanical
Devices Sourcebook
by Neil Sclater / Nicholas Chironis
The fourth edition
of this invention-
inspiring engineering
resource covers the
past, present, and
future of mechanisms
and mechanical
devices. Youll find
drawings and
descriptions of more
than 2,000 compo-
nents that have
proven themselves over time and can
be incorporated into the very latest
mechanical, electromechanical, and mecha-
tronic products and systems. Overviews of
robotics, rapid prototyping, MEMS, and
nanotechnology, along with tutorial chapters
on the basics of mechanisms and motion
control, will bring you up-to-speed quickly
on these cutting-edge topics. $89.95
We accept VISA, MC, AMEX, and DISCOVER
Prices do not include shipping and
may be subject to change.
CNC Robotics
by Geoff Williams
CNC Robotics gives you
step-by-step, illustrated
directions for designing,
constructing, and testing
a fully functional CNC
robot that saves you 80
percent of the price of an
off-the-shelf bot and
that can be customized
to suit your purposes
exactly, because you designed it. Written
by an accomplished workshop bot
designer/builder, this book gives you
everything you need. $34.95
The SERVO Webstore
The Amateur Scientist 3.0
The Complete Collection
by Bright Science, LLC
There are 1,000 projects
on this CD, not to
mention the additional
technical info and
bonus features. It
doesnt matter if youre
a complete novice
looking to do their first
science fair project or a
super tech-head gadget
freak; there are enough
projects on the single CD-ROM to keep you
and 50 of your friends busy for a lifetime!
$26.99
Attention Subscribers ask about your discount on prices marked with an *
N N
E E
W
!
W
!
Forbidden LEGO
by Ulrik Pilegaard / Mike Dooley
Build the Models
Your Parents
Warned You
Against.
Forbidden LEGO
introduces you to
the type of free-
style building that
LEGOs master
builders do for fun
in the back room. Using LEGO bricks in com-
bination with common household materials
(from rubber bands and glue to plastic
spoons and ping-pong balls) along with
some very unorthodox building techniques,
youll learn to create working models that
LEGO would never endorse. $24.95
SERVO 03.2008 75
WebstoreMarch08.qxd 2/6/2008 2:06 PM Page 75
Get your very own limited-edition SERVO Magazine T-shirt. Shirts come
in sizes S, M, L, and are available in either black or white.
All shirts are 100% preshrunk cotton. Please check availability on our website.
SERVO
Magazine
T-Shirts
For men and women
B
lo
B
low
O
u
t
w
O
u
t
S
p
e
c
ia
l
S
p
e
c
ia
l
$
9
.9
5
!
$
9
.9
5
!
From HomoSapien to RoboSapien Before R2D2 there was R1D1
Take This Stuff and Hack It!
by Dave Prochnow
Transform common household items into
really cool stuff. You
don't need to be an
electronics genius to get
started turning everyday
items into high-performing
wonders. With how-to
guru Dave Prochnow's
step-by-step directions
and fully illustrated plans,
even beginners can hack their way to a high-
tech home, cooler toys, and less yard work.
Certain to fire your imagination and start you
plotting new, original, and even more creative
wonders you can make from ordinary house-
hold items, Take This Stuff and Hack It! is the
perfect gift for your inner inventor. $27.95
Building Robots with LEGO
Mindstorms NXT
by Mario Ferrari, Guilio Ferrari
The Ultimate Tool for
MINDSTORMS
Maniacs, the new
MINDSTORMS kit has
been updated to
include a
programming brick,
USB cable, RJ11-like
cables, motors, and
sensors. This book
updates the robotics information to be
compatible with the new set and to show
how sound, sight, touch, and distance issues
are now dealt with. $39.95
NEED MORE
INFO?
Go where
SERVE goes!
The online stor The online store @ e @
www www.ser .ser v vomagazine omagazine.com .com
Robot Builders Bonanza
Third Edition
by Gordon McComb / Myke Predko
Everybodys favorite
amateur robotics book
is bolder and better
than ever and now
features the fields
grand master Myke
Predko as the new
author! Author duo
McComb and Predko
bring their expertise
to this fully-illustrated robotics bible to
enhance the already incomparable content
on how to build and have a universe of
fun with robots. $27.95
SERVO Magazine
Bundles
Published by T & L Publications, Inc.
To order call 1-800-783-4624
Now you can get one years worth of all
your favorite articles from SERVO Magazine
in a convenient bundle of print copies.
Available for years 04, 05, 06, and 07.
$57
per bundle
S
a
v
e
$
1
0
o
ff th
e
n
o
rm
a
l
p
ric
e
!!
Robot Builders Sourcebook
by Gordon McComb
Fascinated by the
world of robotics but
dont know how to
tap into the incredible
amount of informa-
tion available on the
subject? Clueless as
to locating specific
information on robot-
ics? Want the names,
addresses, phone
numbers, and websites of companies that
can supply the exact part, plan, kit, building
material, programming language, operating
system, computer system, or publication
youve been searching for? Turn to Robot
Builders Sourcebook a unique clearing-
house of information that will open 2,500+
new doors and spark almost as many new
ideas. $24.95
76 SERVO 03.2008
BACK ROOM SPECIALS
Concise Encyclopedia
of Robotics
by Stan Gibilisco
Written by anaward-win-
ning electronics author,
the Concise Encyclopedia
of Robotics delivers 400
up-to-date, easy-to-read
definitions that make even
complex concepts
understandable. Over
150 illustrations make the
information accessible at a
glance and extensive cross-referencing
and a comprehensive bibliography facilitate
further research.
$19.95 Sale Price $9.95 Only 2 Left
Pub date: November 12, 2002
SERVO'S CUSTOM
EMBROIDERED HAT
This is one good looking hat, with its stone
washed blue cap and bright yellow/gold
stitching. There's no better way to show
everyone that you're well endowed with a
bigger - than - average brain then by proudly
displaying your affiliation with SERVO
Magazine. $ 14.95*
Hydrogen Powered Racer and Hydrogen Station
Recently named as one of the Best Inventions of 2006 by Time Magazine, the H-racer is now the
best selling fuel cell product in the world. The H-racer is a micro-version of what engineers and
scientists have been dreaming about for real cars: combining hydrogen with oxygen to generate
a DC current to power an electric motor. Unlike a gas-powered car engine, the only byproducts
of this electrochemical process are electricity, heat, and pure water. With the H-racer, you can
witness the power of new energy technology in the palm of your hand. Horizon has created
this unique, patented miniature fuel-cell car and hydrogen refueling station. All you need to do
is add water. Only $115.00 plus S/H
Robotics Demystified
by Edwin Wise
YOU DON'T NEED ARTIFICIAL INTELLIGENCE
TO LEARN ROBOTICS!
Now anyone with an
interest in robotics can
gain a deeper under-
standing without for-
mal training, unlimited
time, or a genius IQ. In
Robotics Demystified,
expert robot builder
and author Edwin Wise
provides an effective and totally painless
way to learn about the technologies used
to build robots!
$19.95
Pub Date October 20, 2004
The Day the Earth
Stood Still
An alien (Klaatu) with his
mighty robot (Gort) land
their spacecraft on Cold
War-era Earth just after
the end of World War II.
They bring an important
message to the planet
that Klaatu wishes to tell
to representatives of all
nations. Directed by: Robert Wise
Actors: Michael Rennie, Patricia Neal
Run Time: 92 minutes $14.95
Forbidden Planet:
50th Anniversary Edition
A starship crew goes to
investigate the silence of a
planets colony only to
find two survivors and a
deadly secret that one of
them has. Dr. Morbius and
his daughter Altaira have
somehow survived a
hideous monster which
roams the planet.
Directed by: Fred M. Wilcox
Run Time: 98 minutes $26.95
Blade Runner
(The Directors Cut) (1982)
In a cyberpunk vision of
the future, man has devel-
oped the technology to
create replicants, human
clones used to serve in the
colonies outside Earth but
with fixed lifespans. In Los
Angeles, 2019, Deckard is a
Blade Runner, a cop who
specializes in terminating replicants.
Directed by: Ridley Scott
Rated: R Violence $19.95
CLASSIC SCIENCE FICTION DVDS!
With this kit, you now can see and
feel the future of energy generation
in your own hands!
Linux Robotics
by D. Jay Newman
If you want your robot
to have more brains than
microcontrollers can
deliver if you want
a truly intelligent,
high-capability robot
everything you need
is right here. Linux
Robotics gives you step-
by-step directions for Zeppo, a
super-smart, single-board-powered robot
that can be built by any hobbyist. You also
get complete instructions for incorporating
Linux single boards into your own unique
robotic designs. No programming experi-
ence is required. This book includes access
to all the downloadable programs you need,
plus complete training in doing original
programming. $34.95 Sale Price $ 29.95
Or order online www.servomagazine.com
SERVO 03.2008 77
W
hen we were at FLOAT the
Floatation Center and Art Gallery
we started thinking about what we
should pick for our December art show,
robots as art came to us in an instant.
Why not fill our gallery with DIY Bay
Area robotic art? Hell, everyone loves
robots, dont they? And isnt it the next
brick and mortar Internet, the job of the
future? Who knew how much work we
were going to be taking on ... I guess
we should be careful what we ask for!
Four furiously busy months later
the Robots are Art art show and
contest came to be. We had 17
robotic artists, including: a robot that
served beer from his belly; a robot
made of toilet paper rolls; a beggar
robot the drove up and down the
sidewalk barking out give me 50
cents; a breathalyzer robot, and lots
of fantastic robotic influenced wall art.
Infiltrating the robotic community
is a bit of an endeavor, but we were
lucky enough to be led to these arti-
sans though a number of connections.
Max Chandler who programs robots to
paint petroglyphs was one of the first
robotic artists to respond to my call for
entries, providing valuable insight into
the community. John Willams, owner
of Boss Robots a local Berkley
robotic toy store, helped spread the
word and Frank Garvey, founder of
Omni Circus a robotic cabaret in San
Francisco was also invaluable. Garvey
eventually led me to David Calkins,
founder of the Robotic Society of
America, who was kind enough to be
one of our robotic art judges, and
helped to spread the word nationally.
Supporting this madness and creat-
ing a tipping point with press was Monty
the gorgeous Anybots robot in his
first open-to-the-public appearance. As
luck would have it, while cold calling Bay
Area robotics companies for sponsorship,
I had the sheer luck to have Scott Wiley
(one of the Anybots engineers) answer
the phone. Wiley just happened to be an
artist. I was soon invited to come meet
Monty and Dexter at the Anybots incuba-
tion facility in Mountain View, CA.
What an amazing group of guys
building peaceful robots of the future!
The team is lead by Trevor Blackwell,
the CEO and founder. He is one of
those brilliant guys that loves to create
the next best thing. Not only did
Anybots decide to sponsor us, they
were bringing Monty to judge the
robotic art. (We did a happy dance on
the drive back to the gallery!)
We had wonderful industry and
non-industry sponsors. The Ugobe
Pleos came to play with the kids at the
show; our Robonova1 was a major hit,
especially since Hitec sent me a Ti Chi
program to entertain the masses. We
had a crowd of hundreds come, with a
line out the door. Our spin the robot
raffle allowed party goers to win hats
and tees from Pandora.com, SERVO
Magazine subscriptions, incredible
wine donated by Mel Knox, wine
broker, and floatation therapy sessions.
The grand prize was an art robot
toy plus a visit to Anybots to meet
Dexter and Monty. Monty autographed
8x10 pictures of himself and handed
them out to the crowd. The local FOX
television station had us on the 10
oclock news. (See robots are fun!)
In addition to wine, beer, and
snacks, the party-goers were able to
pick up free cups, and SERVO
magazines. (Best $5 deal in town!) The
money went to cash prizes for the
winners, who won categories such as
now thats a sexy robot; robots as a
reflection of society; and incorpora-
tion of found objects, to name a few.
During the month long show, we had
800-900 people visit the gallery.
People who never thought they
liked robots are now converts. The
show was a wonderful success, clearly
the art of robotics is a strong draw in
the Bay Area, and we hope that more
venues will take the time to appreciate
robotic artists as much as we do.
Even weeks after the event closed,
we still get people calling about
wanting to see the show. Were
actually still recovering from the first
event! So, be careful what you ask for!
Now were off to build our own xenon
plasma headed robot, and will decide if
Robots are Art 2 is in our future.
Look for our upcoming youtube video
of the show soon! SV
When Art and
Servos Mix
by Allison F. Walton and Filomena Serpa
78 SERVO 03.2008
Appetizer.qxd 2/6/2008 10:34 AM Page 78
I
was talking with my wife, Sue, about
Womens History month and she
asked if an article on Women of
Robotics would be an appropriate
subject for my column. She knew that
I had worked with various women in
my robotics work over the years. After
all, in September of last year I had
written about Bala Krishnamurthy in
People of Robotics. She has been in
the field of robotics for over 25
years and designed and developed
programming languages for Unimates
robots, among many other things.
I never had the opportunity to
meet Bala, just one of many thousands
of women who have specialized in
this field and have made numerous
contributions to the design and science
of robotics. Unfortunately, before
World War II, people used to typecast
men as technical types and women as
homemakers (being experts in the
more gentle fields), but Rosie the
Riveter quickly changed this assump-
tion, for aircraft production and every
other technical field. In my years with
Rockwell, I met and worked with many
extremely talented women in technical
fields who knew more about their spe-
cialty than I could ever hope to know.
I considered NASA astronaut, Dr.
Judith Resnik, who died in the 1986
Challenger Space Shuttle explosion a
good friend as she taught me much of
what I knew at the time about the
Shuttles remote manipulator system. I
was trying to use the RMS arm on my
Space Station robot mobile manipula-
tor/transporter system design and she
was a wealth of practical information
and encouragement on all parts of the
system. The depth of her knowledge in
many fields of science and engineering
just amazed me.
I happened to be at Johnson
Space Center giving a presentation to
NASA management on the mobile
transporter when we stopped for a
moment to view the TV monitor
behind me showing the launch of
Challenger. Needless to say, my
presentation ended with the explosion
of the space shuttle as I had just lost a
friend actually two friends, who
were a great inspiration for my work.
Women in Leadership
of Robotics Groups
Many of the women whom I have
become acquainted with who are
associated with robotics do not work
only within university or industry labs,
but are actively involved with robotics
within the community. Cathy Saxton is
one individual who is one of the driving
forces behind the Seattle Robotics
Society, a group I have been involved
with since moving to Washington in
1998 (Figure 1).
Cathy is presently
the Vice President and
webmaster for the
SRS, but her involve-
ment does not stop
there. She has also
been one of the lead-
ers of SRSs popular
robot workshops that
introduce newcomers
to robot building with
a great robot kit she helped design. She
is also involved with the SRS yearly
Robothon two-day event at Seattle
Center, and is a mentor for nearby
Issaquah High Schools FIRST team.
Jianna Zhang is another woman
who I have come to admire in the field
of robotics (Figure 2). She is an associ-
ate professor of Computer Science at
Western Washington University in
Bellingham, WA. She was the main
supporter in the establishment of the
Bellingham Artificial Intelligence and
Robotics Society (BAIRS) and is the
current President. This is a group that is
open to all in the area who have an
interest in robotics as a hobby, not just
to those in the university.
BAIRS has had several well-attend-
ed exhibits and robot fairs with Jianna
taking the lead in organizing and pro-
moting. I have attended several of the
fairs and meetings, and Jianna and the
society should be proud of the quality
of the different robots presented and
the expertise of the many members.
a
n
d
WOMEN OF ROBOTICS
b y T o m C a r r o l l
FIGURE 1. Cathy Saxton of SRS.
FIGURE 2.
Jianna Zhang
of BAIRS.
SERVO 03.2008 79
Then&Now.qxd 2/6/2008 3:46 PM Page 79
80 SERVO 03.2008
Women at the Helm
of Robotics Industry
and Research
I recently had a very enjoyable
conversation with Helen Greiner (Figure
3) who is the Co-founder and Chairman
of the Board of iRobot Corporation,
maker of the popular Roomba and many
other robotic products for consumers
and the military. At first glance, you
might not identify Helen as the
co-founder of a large technical company.
Youd be wrong if you felt that this petite
and soft-spoken woman would head up
the housecleaning Roomba and Scooba
robots division and leave the Government
and Industrial Robots Division of iRobot
to the other co-founder, Colin Angle.
Greiner overcame her basic shyness
to now speak before thousands of
industrial and government specialists
and start new divisions of their
company. She leaves Angle the tasks of
marketing the highly-successful Roomba
and the consumer half of their corpora-
tion. She has always said, Leaders dont
always look like you think they should.
Under that reserved persona is an
industrial leader who is second to none.
What was the spark that set her off
on this highly successful path? As a
young girl who was born in London, she
grew up on Long Island, NY with an
early interest in science, mathematics,
and engineering. Her mother was a sci-
ence and math teacher who undoubt-
edly had a great influence on her, as did
her father who had a major in chemistry
in college and was a businessman.
From the age of five, she enjoyed a
daily chess match with her father, but
had an even greater interest in her first
RadioShack TRS-80 that she quickly
learned how to program. To her, the
computer was not just another toy with
which to play a few games on and
then set it aside for lack of interest.
Programming this early personal com-
puter was a challenge to her; it gave
her the ability to create something new.
To many of us in the field of robot-
ics, a robot in a story or movie was what
sparked our first interest in these unique
mechanisms we all love. Greiners
epiphany was seeing Star Wars at the
age of 11 and being transfixed by
R2-D2. It was a bit disconcerting to her
to find out that the little robot star of
the film was actually controlled by a
diminutive man (Kenny Baker) inside the
robot. She knew in her heart that real
robots were not too far in the future
and certainly hoped that she may,
someday, be a part of that future. Little
did she realize that she would be one of
the driving forces behind one of the
most successful robot companies.
In high school, she found it quite
natural to do repairs on her Volvo station
wagon. Many girls still find technical
things geeky, but Helen saw no good
reason why women cannot major in
mechanical engineering and similar fields
in college. The salaries are certainly much
higher than most other fields, but, she
says, its so creative and interesting.
Planning for college, she knew that
engineering and computer science
were the courses for her. As she told
me, the proximity to MIT was not only
a plus, but the courses at this
renowned school offered a unique
hands-on atmosphere. She remem-
bered watching the televised competi-
tions between MIT students that were
basically given a bag of parts and told
to construct a robot or similar mechan-
ical device to collect as many ping pong
balls as possible in a given time. This is
the school for me, she thought.
She received a BS in Mechanical
Engineering and a Masters in
Computer Science, but one of the most
important gains from MIT was a close
association with Rodney Brooks and
Colin Angle at MITs Artificial
Intelligence Lab, both of whom would
be instrumental in the future formation
of IS Robotics, later called iRobot.
After graduation, she worked a
few years at NASAs Jet Propulsion
Labs (JPL) in Pasadena, CA. She later
formed California Cybernetics to com-
mercialize some of the unique and use-
ful technical ideas that were developed
at JPL but not used in industry at the
time, and also to perform government-
sponsored research in robotics.
When IS Robotics came into being,
the main interest was walking robots
and artificial creatures. IS Robotics, Inc.,
was formed in 1990 and changed to IS
Robotics Corporation in 1994, then
finally to iRobot Corporation in 2000. In
the early IS Robotics days, DARPA and
the Office of Naval Research were major
customers and ISR developed unique
robots to detect and retrieve mines and
unexploded bombs, swim like fish, and
even crawl up walls. They had an early
contract with the toy company, Hasbro
to develop interactive robots as toys.
With over two million Roomba
floor sweeping robots and many
Scooba floor washing robots sold by
iRobot, Greiner, Angle, and the rest of
their team do not dare rely on their
many laurels, but are actively pursuing
government and industry contracts.
iRobot has brought in Joe Dyer as
President of the Government and
Industrial Robots Division to work with
Helen in this very lucrative market. They
have delivered over 1,200 Packbot
Tactical Mobile Robots to the military
to assist in many facets of on-going mil-
itary operations around the world.
The Packbot can more accurately
be described as a kit as users can con-
figure it in many ways, according to the
particular mission or task. The Packbot
Scout, Explorer, and EOD (explosive
ordinance disposal) or IED (those dead-
ly improvised explosive devices) config-
urations are the more popular uses and
have found to be extremely valuable to
urban SWAT teams and other law
enforcement agencies. There is even a
variation called Fido with bomb sniff-
ing technology for a reasonable $150K.
As iRobots board chair, Greiner
must oversee all divisions of the com-
pany and balance the basic structures
within. iRobot has also developed
strategic alliances with John Deere, the
Clorox Company (Scoobas floor-wash-
FIGURE 3. Helen Greiner of iRobot.
ing fluid), and Taser International. The
new Roomba 530, 560, and 570 greatly-
improved models seem to be a hit.
They no longer get tangled in carpet
fringe, cords, and similar. Priced from
$250 to $400, they are more powerful,
have newer non-contact sensors, a bet-
ter virtual wall, and greater filtration.
The recent Consumer Electronics
Show (CES) in Las Vegas recently
presented iRobot with the Best of
Innovations Design and Engineering
Award for their latest product the
Looj gutter-cleaning robot line priced
from $85-$170 (Figure 4).
I asked Helen why the name and
she told me it reminded her of the
small luge racing sleds as it snakes its
way through a clogged gutter, tossing
the leaves and mess out with its spin-
ning auger. This was iRobots third win
at the CES in four years. Not bad for a
company that didnt exist a decade and
a half ago and is now close to one
quarter billion dollars in annual sales.
I found the whole conversation
with Helen most informative. She
not only speaks with tremendous
enthusiasm about the past accomplish-
ments and future of her company, but
does so with a definite knowledge of
the potential market and how to
address those future customers with a
solid plan of commercialization.
Enthusiasm is one thing, but a solid
technical background coupled with
down-to-earth business sense is what
makes Helen stand out among the
competition. Her forward-looking
instincts and leadership in the industry
have led iRobot to become a world
leader in the robot industry. Helen is
an innovator in technology, government
research, and business, says Rodney
Brooks, her MIT professor and advisor,
member of the board, and mentor to
her and Angle in forming the company.
In 2007 at the ROBOBusiness
Conference, iRobot sponsored an
amateur robot building contest that
required the builder to use their iRobot
Create platform as a starting point
(Figure 5). Danh Trinh won the $5,000
prize with his teleoperated Personal
Home Robot that can water plants,
control home systems, and act as a
media center (Figure 6).
Helen has led the company into
winning an Army $51.4M con-
tract for a Small Unmanned
Ground Vehicle (SUGV), and is
continuing development of
the AWARE Robot Intelligence
Systems and improvements in
the iRobot Create platform.
She was recently named by
the Kennedy School at Harvard as one
of Americas Best Leaders and was
honored with the Pioneer Award from
the Association for Unmanned Vehicle
Systems International (AUVSI) in appre-
ciation for her work in military robotics.
She was also recognized by
Fortune Magazine as one of its Top
10 Innovators of 2003 and named the
Ernst and Young New England
Entrepreneur of the Year with co-
founder Angle. Earlier, she was named
an Innovator for the Next Century by
Technology Review Magazine in 1999.
She has appeared in many prestigious
magazines and TV shows, sits on sever-
al scientific boards, and is a frequent
speaker at many technical symposiums.
Helen certainly does not spend all
her time thinking of robots, however.
She is an avid snowboarder (she loves
the speed), scuba-diver, rock climber,
paint-baller, ice hockey player, and
wind surfer, yet, still has time to tend
her English Garden. I doubt that shes
ever going to slow down in the near
future. There are too many potential
robotic applications out there in the
world for her to conquer.
Cynthia Breazeal
Cynthia Breazeal is an Associate
Professor in MITs Media Lab and is
Director of the Robotic Life
Group (Figure 7). I didnt
plan on highlighting only
women from MIT but I felt
that these two women
(Greiner and Breazeal) had
taken dramatically different
directions after graduation,
yet possess many similari-
ties. Many first-rate
female robotics researchers
have come from Carnegie
Mellon, Stanford, and
other premier universities
around the US and world,
and are making their
names known amongst their peers.
These two women know each other
quite well, but Cynthia felt that her place
was in academia, imparting her knowl-
edge on others and doing basic research.
Cynthia did not start out as a robot
enthusiast as a young girl; she was more
interested in climbing tall trees with her
brother, Bill, plus swimming, tennis, soc-
cer, and track. It was a trip to the movies
as a 10-year old Star Wars in fact
that led her to a love for R2-D2 and
C-3PO and robots in general, much like
iRobots Greiner. Sports continued to be
her first love as she felt that these robots
with such admirable personalities were
impossible to build. Her desire was to be
a medical doctor, though she grew up in
a family that was science and technolo-
gy oriented. It was her parents who
encouraged her to major in engineering.
Breazeal received her BS in
Electrical and Computer Engineering
from UC Santa Barbara in 1989, and
her Masters (1993) and Doctorate
(2000) from MIT in Electrical
Engineering and Computer Science.
UCSB got her hooked on engineering,
physics, chemistry, and astronomy
paving the way for her to ultimately
develop the type of robot she wanted.
She worked at the UCSB robotics
center part time during her summers to
get her hands dirty, as she says. She
FIGURE 4. Side view of Looj Gutter-Cleaning Robot.
FIGURE 5. Helen holding the
Create Platform at the CES Show.
FIGURE 6.
iRobots Create
contest winner.
SERVO 03.2008 81
actually seriously considered being an
astronaut at this time, so space robotics
seemed to be an ideal path towards this
goal. She loved learning how things
worked and solving mechanical problems.
It was MITs amazing way of presenting
engineering to students, plus Professor
Rodney Brooks, that gave Cynthia the
freedom to delve deeply into highly
expressive anthropomorphic robots
basically, robots with personalities.
The social responses she observed
from Star Wars C-3PO were scripted
acting, but were the inspiration for her
to delve much deeper into robot social
interaction. In 2003, she married
Bobby Blumofe, one of comedian
Jack Bennys grandsons, and has two
children at home.
Cynthia has
been studying and
building autonomous
robots for well over a
decade now from
the insect-like robotic
crawlers that made
MIT and the grad
students famous,
to her expressive
anthropomorphic (human like) robots
such as the upper torso robot, Cog, and
the sociable robot, Kismet. Her first book,
Designing Sociable Robots in 2002, and
co-written Biologically Inspired Intelligent
Robots in 2003 have been valuable
references for those following her work in
these unique types of robots. These are
not just simple robots that only respond
to one type of stimulus with basic
outputs; these are complex social robots
who learn and respond directly to people
in their environments. She is more inter-
ested in finding out how robot can be
more like us humans rather that having us
adapt to their particular requirements.
Figure 8 shows Kismet Breazeals
crowning achievement. As you walk
into her Artificial Intelligence Lab at
MIT, you will probably come face to face
literally with this amazing artificial
creature. Its termed anthropomorphic
as it is somewhat human-sized and
shaped, but it reminds me more of
Yoda from Star Wars. It will probably
look up at you, its droopy ears perking
up, its large eyes suddenly seeing you.
The edges of its rubber lips will move
upward into a smile as its eyebrows
spring upward. You cannot help from
smiling back; I know Kismet is a
machine but it likes me! You look
around to see if anybody is watching
you. Yes, Cynthia is across the room
smiling; her creation has accomplished
its objective it has interacted with yet
another human (see Figure 9).
She hasnt stopped her work with
this robot. She sees many more steps to
making robots truly useful around
humans. Japan, and even the US, are
finding a rapidly growing elderly popula-
tion that will soon require basic help in
independent living. A robot that has a
true personality a pleasant personali-
ty will certainly be more acceptable to
a technology-resistant elderly person
than a metallic, medical-looking appli-
ance. Breazeal, Greiner, and many other
robotics researchers are closing in on
that day when anthropomorphic robots
will be commonplace in our homes.
In Conclusion
All of the women who Ive written
about in this article will certainly agree
that there are still social barriers to
women pursuing science and engineer-
ing fields. They will also all agree that
the fact that they are a woman does
NOT bring anything special to the field
of robotics, or any field for that matter.
It is not what they are but who they are.
Getting the education that you
want is the key for anyone even the
slightest bit interested in fields of sci-
ence or technology. Yes, these women
excel in their fields and specialties, but
many tens of thousands of women are
also contributing to the advancement
of robotics in this country. SV
FIGURE 8. Cynthias Kismet.
FIGURE 7.
Cynthia
Breazeal of MIT.
FIGURE 9. Cynthia and Kismet.
Tom Carroll can be reached via email
at TWCarroll@aol.com
CONTACT THE AUTHOR
82 SERVO 03.2008
All Electronics Corp. ..........................27, 66
AP Circuits/e-pcb.com ............................64
AWIT ..........................................................66
Boca Bearings .......................................7, 66
CipherLinx Technologies .........................66
CrustCrawler .............................................83
Electronics123 ..........................................27
Futurlec .....................................................66
Hitec ..........................................................61
Images Co. ................................................66
Jameco ......................................................18
Lorax Works ........................................27, 66
Lynxmotion, Inc. .........................................3
Maxbotix ...................................................66
Net Media ...................................................2
Online-tech-training.com .........................66
Parallax, Inc. ...............................Back Cover
PCB Pool ..............................................66, 71
Pololu Robotics & Electronics ..........57, 66
RCG Research ...........................................64
ROBOBusiness ..........................................13
RoboGames ..............................................37
Robotis Co. Ltd. .......................................70
Robot Power ...............................................7
RobotShop, Inc. .................................65, 66
Schmartboard...........................................27
Shooting Star Technology..................12, 66
Solarbotics/HVW .....................................71
Technological Arts ...................................66
Trossen Robotics ......................................25
Vantec .......................................................12
Advertiser Index
backcvr.qxd 2/4/2008 10:02 AM Page 84

Servo2008 03

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Servo2008 03

Uploaded by

Copyright:

Available Formats

V

You might also like