Sapienza Universit

`
a di Roma Facolt
`
a di Ingegneria
DIS - Dipartimento di Informatica e Sistemistica Antonio Ruberti
Nonlinear Robust Control
of an Highly Manoeuvrable
Missile
supervisor: candidate:
Chiar.mo Prof. Salvatore Monaco Giovanni Mattei
co-supervisor:
Ing. Emidio Pizzingrilli
A.Y. 2010-2011
Introduction
Automatics and control theory should be considered at the top of modern engineering
science. Control systems are everywhere and, over time, they are going to be more and
more essential (not just incidental, as happens now) for the proper functioning of any
system.
This thesis is about control system design for a generic air-to-air missile autopilot and
guidance system. For the autopilot, a nonlinear recursive approach is proposed, beginning
from a simpler, nominal, implementation to arrive, in the end, to a more complicated
robust one. The rst two control systems proposed are based on a wiser version of
Backstepping, which uses a non-quadratic Lyapunov function to achieve in a more natural
way some necessary dynamic cancellations. A faster version of these two autopilots is
also designed, including nonlinear damping terms. Then, nonlinear robust control theory
is introduced, highlighting the structure of the generic nonlinear uncertain system. A
nonlinear robust roll autopilot is designed, including in the robust recursive approach, that
shall represent the baseline for the robust versions of the autopilot, also a revisited version
of Lyapunov redesign, which has no counterpart in control theory literature. For pitch and
yaw dynamics are, nally, proposed the static and dynamic robust recursive designs, two
rigorous and systematic approaches to the problem of controlling a nonlinear uncertain
system, which seem to be well suited to handle the structure of the missile vector-elds.
To derive the equations of all the nonlinear controllers, a complete, nonlinear dynamical
model of the missile is necessary: this is very important, because nonlinear control
represents a model based design methodology. The simulation model was developed in the
rst part of the work, together with the simplied models used for control design. Also, an
accurate and very physical-oriented open-loop analysis of the missile behaviour is presented,
with some simulations regarding the statically unstable dynamics, the phenomenon of
peaking related to nite time explosions and the non-minimum phase.
The nal part of the work is dedicated to the guidance system design. A nonlinear,
Lyapunov-based, approach is used to guarantee the nite-time convergence to zero of
the line-of-sight angular rates. The guidance loop is closed around the dynamic robust
recursive controller, which in simulation will show the best performance. In this way, the
guidance system and the autopilot will be both of energetic and nonlinear nature, and
9
INTRODUCTION 10
their integration and coupled functioning shall be really good, as simulations will show in
the last part of the thesis.
Part I
Overview on rocket science
11
Chapter 1
A brief history
As a presentation to this work, it seems convenient to introduce the reader to the world of
rocket science by means of its history, its development as a stronghold for modern warfare
and its pretty recent employment in space exploration. This part of the work is freely
taken by [28].
Rockets have been used as early as A.D. 1232, when the Chinese employed them as
unguided missiles to repel the Mongol besiegers of the city of Pein-King (Peiping). Also, in
the fteenth century, Korea developed the sinkijon
1
(or Sin-Gi-Jeon) rocket. Manufactured
from the early fteenth to mid-sixteenth century, the sinkijon was actively deployed in the
northern frontiers, playing a pivotal role in fending o invasions on numerous occasions.
Once out of the rocket launcher, the re-arrows were set to detonate automatically near
the target area. Also, the high-powered rearm was utilized in the southern provinces to
thwart the Japanese marauders. The main body of the sinkijons rocket launcher was ve
to six meters long, the largest of its kind at that time. A sinkijon was capable of ring
as many as one hundred re-arrows or explosive grenades. The re-arrow contained a
device equipped with gunpowder and shrapnel, timed to explode near the target. The
introduction of gunpowder made possible the use of cannon and muskets that could throw
projectiles at great distances and with high velocities. It was desirable (in so far as the
study of cannon re is desirable) to learn the paths of these projectiles, their range, the
heights they could reach, and the eect of muzzle velocity. Several years later, the sinkijon
went through another signicant upgrade, which enabled it to hurl a re-arrow made up
of small warheads and programmed to detonate and shower multiple explosions around
the enemy. In 1451, King Munjong ordered a drastic upgrade of the hwacha (a rocket
launcher on a cartwheel). This improvement allowed as many as one hundred sinkijons to
be mounted on the hwacha, boosting the overall repower and mobility of the rocket.
Since those early times and in one form or another, rockets have been used as weapons
and machines of war, for amusement through their colorful aerial bursts, as life-saving
1
Sinkijon means ghost-like arrow machine.
12
CHAPTER 1. A BRIEF HISTORY 13
equipment, and for communications or signals. The lack of suitable guidance and control
systems may have accounted for the rockets slow improvement over the years. Strangely
enough, it was the airplane rather than the rocket that stimulated the development of a
guided missile as it is known today.
In the twentieth century, the idea of using guided missiles came during World War I.
Specically, and as stated above, the use of the airplane as a military weapon gave rise
to the idea of using remote-controlled aircraft to bomb targets. As early as 1913, Rene
Lorin, a French engineer, proposed and patented the idea for a ramjet powerplant. In
1924, funds were allocated in the United States to develop a missile using radio control.
Many moderately successful ights were made during the 1920s with this control, but by
1932 the project was closed because of lack of funds. Radio-controlled target planes were
the rst airborne remote-controlled aircraft used by the Army and Navy.
Dr. Robert H. Goddard was largely responsible for the interest in rockets back in
the 1920s. Early in his experiments he found that solid-propellant rockets would not
give him the high power or duration of power needed for a dependable supersonic motor
capable of extreme altitudes. On March 16, 1926, Dr. Goddard successfully red the
rst liquid-propellant rocket, which attained an altitude of 56 m and a speed of 97 km/h.
Later, Dr. Goddard was the rst to re a rocket that reached a speed faster than the
speed of sound. Moreover, he was the rst to develop a gyroscopic steering apparatus for
rockets, rst to use vanes in the jet stream for rocket stabilization during the initial phase
of a rocket in ight, and the rst to patent the idea of step rockets.
The rst ight of a liquid-propellant rocket in Europe occurred in Germany on 14
March 1931. In 1932 Captain Walter Dornberger (later a general) of the German Army
obtained the necessary approval to develop liquid-propellant rockets for military purposes.
Subsequently, by 1936 Germany decided to make research and development of guided
missiles a major project, known as the Peenem unde Project , at Peenem unde, Germany.
The German developments in the eld of guided missiles during World War II were the
most advanced of their time. Their most widely known missiles were the V-1 and V-2
surface-to-air missiles (note that the designation V1 and/or V2 is also found in the
literature). As early as the spring of 1942, the original V-1 had been developed and
ight-tested at Peenem unde.
In essence, then, modern weapon (missile) guidance technology can be said to have
originated during World War II in Germany with the development of the V-1 and V-2
(German: A-4; the A-4 stands for Aggregat-4, or fourth model in the development type
series; the V stands for Vergeltungswae, or retaliation weapon, while some authors claim
that initially, it stood for Versuchsmuster or experimental model ) surface-to-surface missiles
by a group of engineers and scientists at Peenem unde. In the spring of 1942 the original
V-1 (also known by various names such as buzz bomb, robot bomb, ying bomb, air
CHAPTER 1. A BRIEF HISTORY 14
torpedo, or Fieseler Fi-103) had been developed and ight-tested at Peenem unde. Thus,
the V-1 and V-2 ushered in a new type of warfare employing remote bombing by pilotless
weapons launched over a hundred miles away through all kinds of weather, day and night.
The V-1 (Fig. 1.1) was a small, midwing, pilotless monoplane, lacking ailerons but
using conventional airframe and tail construction, having an overall length of 7.9 m and a
wingspan of 5.3 m. It weighed 2180 kg, including gasoline fuel and an 850 kg warhead.
Powered by a pulsejet engine and launched from an inclined concrete ramp 45.72 m long
and 4.88 m above the ground at the highest end, the V-1 ew a preset distance, and then
switched on a release system, which deected the elevators, diving the missile straight
into the ground. The engine was capable of propelling the V-1 724 km/h. A speed of 322
km/h had to be reached before the V-1 propulsion unit could maintain the missile in ight.
The range of the V-1 was 370 km. Guidance was accomplished by an autopilot along a
preset path. Specically, the planes (or missiles) course stabilization was maintained by a
magnetically controlled gyroscope that directed a tail rudder. When the predetermined
distance was reached, as mentioned above, a servomechanism depressed the elevators,
sending the plane into a steep dive. The V-1 was not accurate, and it was susceptible to
destruction by anti-aircraft re and aircraft.
Figure 1.1: A model of V-1.
The V-2 (Fig. 1.2) (A-4) rocket was one of the most fearsome weapons of WWII.
Successor to the V-1 buzz bomb, the V-2 inicted death, destruction, and psychological
fear on the citizens of Great Britain. In essence, the V-2 was the rst long-range rocket-
propelled missile to be put into combat. Moreover, the V-2 was a liquid-propellant, 14
m rocket that was developed between 1938 and 1942 under the technical direction of Dr.
Wernher von Braun and Dr. Walter Dornberger, Commanding General of the Peenem unde
Rocket Research Institute. Since from the beginning of WWII, Germany was highly
interested in American physicist Robert H. Goddards research. Before 1939, German
CHAPTER 1. A BRIEF HISTORY 15
scientists occasionally contacted Goddard directly with technical questions. Wernher von
Braun used Goddards plans from various journals and incorporated them into the building
of the Aggregat (A) series of rockets, as we will see. In 1963, von Braun reected on the
history of rocketry, and said of Goddards work: His rockets ... may have been rather
crude by present-day standards, but they blazed the trail and incorporated many features
used in our most modern rockets and space vehicles.. The V-2 had movable vanes on the
outer tips of its ns. These ns were used for guidance and control when the missile was in
the atmosphere, which would be for most of its ight when used as a ballistic weapon. It
also had movable solid carbon vanes projecting into the rocket blast for the same purpose
when it was in raried atmosphere. The rst V-2, which landed in England in September
1944, was a supersonic rocket-propelled missile launched vertically and then automatically
tilted to a 41 deg 47 deg angle a short time after launch. Furthermore, the V-2 had a
lifto weight of 12873 kg, developing a thrust of 27125 kg, a maximum acceleration of
6.4g, reaching a maximum speed of about 5705 km/h, an eective range of about 354 km,
carrying a warhead of 998 kg. A at-Earth model was assumed. Like the V-1, the V-2
was not known for its accuracy. For instance, the V-2 had a dispersion at the target of
16 km over a range of 322 km. Active countermeasures against the V-2 were impossible
at that time. Except for its initial programmed turn, it operated as a free projectile at
extremely high velocity. The V-2 consisted of two main parts:
1. a directional reference made up of a gyroscopic assembly to control the attitude of
the missile and a clock-driven pitch programmer;
2. an integrating accelerometer in order to sense accelerations along the thrust axis of
the missile, thereby determining velocity, and to cut o the engine upon reaching a
predetermined velocity.
In essence, the V-2 system was the rst primitive example of inertial guidance, making
use of gyroscopes and accelerometers.
CHAPTER 1. A BRIEF HISTORY 16
Figure 1.2: An exemplar of V-2.
Several other German missiles were also highly developed during World War II and
were in various stages of test. One of these, the Rheinbote (tr. Rhein Messenger), was also
a surface-to-surface missile. This rocket was a three-stage device with booster-assisted
takeo. Another anti-aircraft or surface-to-air missile, the Wasserfall (tr. Waterfall), was
a remote radio-controlled supersonic rocket, similar to the V-2 in general principles of
operation (e.g., both were launched vertically). It had an infrared proximity fuze and
homing device for control on nal approach to the target and for detonating the warhead
at the most advantageous point in the approach. Propulsion was to be obtained from
a liquid-propellant power plant, with nitrogen-pressurized tanks. Another surface-to-air
missile, the Schmetterling (tr. Buttery), designated HS-117, was still in the development
stage at the close of the war. The Enzian was another German surface-to-air missile
(SAM). Designed to carry payloads of explosives up to 453.6 kg, it was intended to be used
against heavy-bomber formations. In addition to the SAMs Germany had developed an
air-to-air missile, designated the X-4. The X-4 was designed to be launched from ghter
aircraft. Propelled by a liquid-propellant rocket, it was stabilized by four ns placed
symmetrically. Guidance was accomplished by electrical impulses transmitted through a
CHAPTER 1. A BRIEF HISTORY 17
pair of ne wires from the ghter aircraft. This missile was claimed to have been own,
but it was never used in combat.
A project for developing missiles in the U.S.A. during World War II was started in
1941. In that year the Army Air Corps asked the National Defense Research Committee
to undertake a project for the development of a vertical, controllable bomb. At the close
of World War II the Americans obtained sucient components to assemble two to three
hundred V-2s from the underground factory, the Mittelwerk, near Nordhausen, Germany.
The purpose of this was to use these V-2s as upper-atmosphere research vehicles carrying
scientic experiments from JPL (Jet Propulsion Laboratory), Johns Hopkins, and other
organizations. In essence, the ballistic missile program in this country culminated with
the development of the Atlas ICBM (intercontinental ballistic missile), with the presence
of Von Braun, who entered U.S.A. at the end of WWII and became a naturalized U.S.
citizen in 1955. In October 1953, and under a study contract from the U.S. Air Force, the
Ramo-Woolridge Corporation (later Thomson-Ramo-Woolridge, or TRW) began work on a
new ICBM. Within a year the program passed from top Air Force priority to top national
priority. The rst successful ight of a Series A Atlas ICBM took place on December
17, 1957, four months after the Soviet Union had announced that it also had an ICBM.
By the mid-1959, more than eighty thousand engineers and technicians had participated
in this program. Wernher Von Braun was the director of the Development Operations
Division of the Army Ballistic Missile Agency (ABMA), in which a big number of german
scientists and physicists worked to contribute to the future of U.S.A. Defense System.
In a paradoxical way, one could say that U.S.A. won the Cold War thanks rst of all to
german know-how. The great Von Braun knowledge on propulsion and ight dynamics
were nally exploited in an amazing manner by NASA, under which he served as director
of the newly-formed Marshall Space Flight Center and as the chief architect of the Saturn
V (Fig. 1.3) launch vehicle, the super-booster that propelled the Apollo spacecraft to the
Moon.
CHAPTER 1. A BRIEF HISTORY 18
Figure 1.3: The Saturn V, a multi-stage liquid-fueled booster (launch vehicle) used by
NASAs Apollo and Skylab programs from 1967 until 1973.
According to one NASA source, he is without doubt, the greatest rocket scientist in
history. His crowning achievement [...] was to lead the development of the Saturn V
booster rocket that helped land the rst men on the Moon in July 1969.. He received the
1975 National Medal of Science for his contribution to space exploration.
Chapter 2
State of the art and system design
The aim of this chapter is to introduce the reader deeper into the world of missiles,
by subdividing them into categories, describing briey their ight envelope and shortly
introducing all of the systems that participate in the accomplishment of a generic mission,
such as, for instance, interception. Finally, an outcome of the design process of a generic
missile is depicted, with a particular reference to the development process of the missile
considered in this work. Sources of inspiration for this Chapter and its gures are [28]
and [6].
2.1 Missile categories
Strictly speaking, missiles can be divided into two categories:
1. guided missiles (also called guided munitions), or tactical missiles
2. unguided missiles, or strategic missiles.
Guided and unguided missiles can be dened in this way:
Guided Missiles. To the guided class of missiles belong the aerodynamic guided
missiles. These are missiles that use aerodynamic lift to control their direction of
ight. An aerodynamic guided missile can be dened as an aerospace vehicle, with
varying guidance
1
capabilities, that is self-propelled through the atmosphere for
the purpose of inicting damage on a designated target. Stated another way, an
aerodynamic guided missile is one that has a winged conguration and is usually
red in a direction approximately towards a designated target and subsequently
receives steering commands from the ground guidance system (or its own, on board,
guidance system) to improve its accuracy. Guided missiles may either home to the
1
The word guidance is to be intended as the means by which a missile steers to, or is steered to, a
target. In guided missiles, missile guidance occurs after launch.
19
CHAPTER 2. STATE OF THE ART AND SYSTEM DESIGN 20
target (on-line target acquisition and path planning), or follow a non-homing preset
course (simple trajectory tracking, o-line path planning). Homing missiles may
be active, semi-active, or passive. Non-homing guided missiles are either inertially
guided or preprogrammed.
Unguided Missiles. Unguided missiles, which includes ballistic missiles, follow the
natural laws of motion under gravity to establish a ballistic trajectory. Examples
of unguided missiles are Honest John, Little John, and many artillery-type rockets.
Note that an unguided missile is usually called a rocket and is normally not a threat
to airborne aircraft.
Typically, guided missiles are homing missiles, which include the following:
1. a propulsion system;
2. a warhead section;
3. a guidance system;
4. a control system or autopilot;
5. one or more sensors (e.g., radar, infrared, electro-optical, lasers), which are part
of the navigation system. This is in some literature overlapped to the guidance &
control systems and represent also the on-line mapping/planning capability of the
missile.
Movable control surfaces (wings, canards or others) are deected by commands from the
guidance system in order to direct the missile in ight; that is, the guidance system will
place the missile on the proper trajectory to intercept the target. The way these surfaces
have to move to achieve the acceleration commands elaborated by the guidance systems
represents the spirit of the control system of the missile, the autopilot.
Homing guidance may be, as already said, of the active, semi-active, or passive type.
Active guidance missiles are able to guide themselves independently after launch to the
target. These missiles are of the so-called launch-and-leave class. Therefore, an active
guided missile carries the radiation source on board the missile. The radiation from the
interceptor missile is radiated, strikes the target, and is reected back to the missile.
Thus, the missile guides itself on this reected radiation. A semi-active missile uses a
combination of active and passive guidance. A source of radiation is part of the system, but
is not carried in the missile; that is, it is dependent on o-board equipment for guidance
commands. More specically, in semi-active missiles the source of radiation, which is
usually at the launch point, radiates energy to the target, whereby the energy is reected
back to the missile. As a result, the missile senses the reected radiation and homes on it.
CHAPTER 2. STATE OF THE ART AND SYSTEM DESIGN 21
A passive missile utilizes radiation originated by the target, or by some other source not
part of the overall weapon system. Typically, this radiation is in the infrared region, but
other sources are frequently used and competitive. In Fig. 2.1 these three congurations
of the guidance system are illustrated.
Figure 2.1: Homing missile guidance types. (Originally published in The Fundamentals
of Aircraft Combat Survivability Analysis and Design, R.E. Ball, AIAA Education Series,
c 1985.)
Non-homing guided missiles have to accomplish missions which do not require an
on-line update of the guidance law, so their path is pre-determined before the launch.
They eventually have to use inertial navigation to adjust errors that deviate them from
the desired trajectory, which is frequently chosen in dependence of optimality criteria.
A slightly dierent kind of non-homing missile can be identied in the presence of an
adaptation law of the pre-set path, with the aim to render it optimal for the particular
kind of mission, also if previously badly perturbed by disturbances. So, as a matter of
fact, this type of adaptive guidance set this genre of missiles between the two classes of
tactical weapons already depicted.
Ballistic missiles belong to the strategic missile class, and are characterized by their
trajectory. A ballistic missile trajectory is composed of three parts.
1. the powered ight portion, which lasts from launch to thrust cut-o (or burnout);
2. the free-ight portion, which constitutes most of the trajectory and is governed by
the laws of motion under gravity (gravity turn, parabolic trajectory);
CHAPTER 2. STATE OF THE ART AND SYSTEM DESIGN 22
3. the re-entry portion, which begins at some point (not always precisely denable),
where the atmospheric drag becomes a signicant force in determining the missiles
path, and lasts until impact on the surface of the Earth (i.e., a target).
Typically, ballistic missiles rely on one or more boosters and an initial steering vector.
They are typically launched in vertical position. After a roll manoeuvre, made to gain
a correct attitude, a pitch manoeuvre is made to bend the path of the rocket. Once in
ight, they maintain their attitude steering vector with the aid of gyroscopes. Therefore,
a ballistic missile may be dened as a missile that is guided during the powered portion of
the ight by deecting the thrust vector, becoming a free-falling body after engine cut-o,
following a so-said gravity turn parabola. However, as already noted, in ballistic missiles
part of the guidance occurs before launch, where an optimal or quite-optimal trajectory is
computed, often solving a two point boundary value problem
2
. Hence, pre-launch errors
translate directly into miss distance, if feedback control is not introduced in the missile
system
3
. In fact, one important feature of these missiles is that they are roll stabilized,
resulting in simplication of the analysis and synthesis of the control system, since there
is no coupling between the longitudinal and the lateral modes. As a result, autopilots for
this kind of missile are of primary importance. Ballistic missiles are the type least likely
to be intercepted and, also, they can have surprising accuracy. Ballistic missiles can be
classied according to their range: short (up to 600 km), intermediate (from 600 km to
4700 km) or long (over 4700 km).
Essentially, the dierence between the ballistic (strategic) and aerodynamic (tactical)
missiles lies in the fact that the former does not rely upon aerodynamic surfaces to produce
lift and consequently follows a ballistic trajectory when thrust is terminated. Aerodynamic
missiles, as stated earlier, do have a winged conguration, and the trajectories that they
can achieve are much more complicated and less predictable. Also, in ballistic missiles the
intent is to hit a given map reference while in aerodynamic missiles it is to intercept a
moving and at times highly manoeuvrable target. In Figure 2.2 a typical subdivision in
categories of missiles is depicted.
2
A set of Partial Dierential Equations to be solved, which have to satisfy initial and nal value
constraints.
3
The typical actuation system utilized for feedback control is Thrust Vector Control (abbr. TVC), with
the presence of little jet engines on the fuselage sides for roll regulation.