You are on page 1of 4

CrashAlerts App: Protecting Individuals From

Direct Collisions During Mobile Interaction


Farhan Azam Memon
Department Of Computer Science
Shaheed Zulfiqar Ali Bhutto
Insititute Of Science and Technology
Hyderabad
Email: farhanmemo@gmail.com

Abstract the user by beeping loudly along with the “RED


ALERT” on individual’s mobile screen so that they
We developed a CrashAlerts app through which human shall stop walking immediately or head-up to
individuals can receive crash alerts on the go if they see what is it and somehow get rid of colliding with
are about to collide into something that could be that object.
harmful for them. We observed the environment
randomly and then took survey to find out what
people think about eyes busy interaction with the
2. Literature Review
mobile phone device while walking on the roads or 2.1 Related Work
anywhere. Then we designed and implemented the “As the technology is moving fast ahead, interaction
application with various resources. The outcome was with the mobile phone devices and the other type of
better that people loved the concept when they were devices has been increased and is being used during
asked to perform the task and participate in the the in-process of other tasks. This needs to take
experiment. It will help people to walk and interact serious attention towards the dangers that
with the device concurrently in a better and safer individuals might face.” [1]
way that there is minimized chance of collision that
may cause in dangerous accident suddenly. This “Using mobile phone devices while walking and
concept can be taken further on in future with better keeping eyes-busy interaction with the devices is a
findings as the technology is moving fast and will be leading danger that is life threatening for sure. It
improving with time. requires attention to make it safer as possible for the
individuals who are very much in interaction with
1. Introduction the devices while even walking on the roads.
CrashAlerts might help them walk with more safety
We have seen how individuals are busy these days and precautions must be taken”[2]
with mobile interaction while walking on roads and
the corridors of their universities that they even do “Pedestrian safety must be taken into consideration
not see what is coming. Their eyes busy interaction as anything that distracts pedestrian concentration
can sometimes be really dangerous for them as they towards the walking is been a cause of fatal and
can collide with objects and collapse. In an study, grave accidents. So, Pedestrian Distraction Safety
researchers found out that their shall be a Framework shall be taken into implementation to
mechanism that interacts with the individuals on help.[3]
their mobile screen with a sounding beep to alert “Texting and walking is the subject to hundreds of
them of collisions with any objects in their way or deaths every year. It is recommended by the
even humans that are walking along or crossing by, National Safety Council to hold your device up
and even the vehicles on the roads. higher if it is very much important to attend the call
We introduce you an application that is connected to or read the incoming texts, to see if there is any
a device that processes the images as fast as possibly danger towards you. In fact, use text-reading and
can and keep track of any objects that comes in voice recognition software so that you can speak on
individual’s way and if there is any detected, it alerts
the go but your eyes shall be on the road to avoid time if they’re about to crash. 100% of people
any danger and walk safely.”[4] agreed that it is helpful to have such an mechanism
in order to protect them from colliding.
“It is much important to know what measures shall
be taken to avoid direct collisions with the objects 3.2 Design
and have much safer life to live even if it is
Based on the observations and some previous
important to text and walk at the same time. But
research, we created the Alerts Application that will
avoiding it is the best and final way you can possibly
by default be running in the background of the
be recommended to.”[5]
Android Mobile Phone of the individual that will
“OpenCV is the best Open Source Computer Vision automatically adjust the Rotating Camera itself
Library that is used for image processing and has towards the forward direction in which the
been created with many image processing individual is walking (generally it shall get set to 90
algorithms to help identify the objects through the Degrees in forward direction). Additionally, there
camera and visualize them to identify them,”[6] will be an option of Manually adjusting the camera
towards the forward direction. And the camera shall
2.2 Contribution be set in such a way that it can capture the objects
By taking the related work into consideration, we from the ground to the top. The Default Camera is
have found out how it can help to enhance the kept in working as well. Every second, images will
usability and safety for the people who are much be processing towards the Image Processing App
more interacted with their mobile phones. So, we through the Bluetooth connectivity and recognizing
made observations, researched on it and found out the objects and obstacles in the way while the person
that the best possible way to accomplish it and make walks. If something comes in the way that the user
the environment safer for the pedestrians. So, we hasn’t noticed for 5 meters. Then the beeping sound
implemented the CrashAlerts App concept to protect will start buzzing on the device. Then if the object is
individuals from direct collisions. just 2 meters far, then the “ALERT: STOP
IMMEDIATELY AND WATCH AHEAD” message
3. Methodology will appear on the user’s screen with loud beeping
sound that the person couldn’t resist. Then if the
3.1 Observations person has stopped, the beep will stop buzzing and
So, we took observations through multiple sources. the message will automatically disappear, or if
First, we observed some young-aged people (e.g doesn’t in some cases, the user can manually close
students) randomly in the university corridors and the message. The person could configure other
on the streets that keep an eyes-busy interaction with settings too, such as, setting beeping volume, enable/
the smartphone and walking was bringing up the disable the vibration, enable/disable blinking of
close-calls with the objects and were alerting screen, setting range of early alerts (up to 15
themselves through natural human alertness like if meters). The app will be having a simple User
they’re about to crash. Some of them we saw even Interface that shall not be difficult to configure and
collided and some of them stopped immediately use for individuals of any age. We tried to design it
before colliding into each other (as their eye sight in such a way to avoid any complexities that put
gave natural alertness). But it was seemingly very pressure on the processors or at the user-end in terms
dangerous for the individuals out there. Then we of understanding. At the other hand, the image
also took a survey from a number of people about if processing app will be made easier to configure as
they feel it is dangerous to use smartphone while well. A user can set fps (frames per second) from
walking and most of them were highly concerned low quality to best quality, depending on the
about it (more than 70%). They even admitted that processor of the device that the device shall not lag
they have collided to something while using the or hang during the use in order to avoid
smartphone and walking concurrently that they lose inconvenience. Otherwise, the app is all made
their focus. We then coming up with an idea, also simple for users to understand, configure and use.
asked them if it is helpful to have an application
mechanism on their devices that alerts them right on
3.3 Implementation 4. Results
We created an mobile application that recognizes the So the experiment was for few minutes in which the
nearby objects that comes in the way of direction in individuals spoke about how it helped them to sense
which the individual is walking. We developed it in if they’re about to dodge into each other. Crash
C# Programming Language that for recognizing Alerts individuals were able to perform simpler
the objects and alerts the user for any obstacles corrective actions (e.g dodging and slowing down)
coming in its way. The application will run on to avoid any possible collisions, which provided
devices with Android version greater than 5.0. individuals additional time and space for other more
And then there shall be another mobile application complex corrections (e.g heads-up and full stops)
that does image processing concurrently and and though, it leads to safer walking than before.
possibly it can be done by using OpenCV (but
OpenCV is used on both applications; the main and Individuals discussed about there shall be alerts
the one for background image processing). Then based on the type, position, direction and speed of
there shall be a Movable Camera (a camera that the object and then calculate the danger level of the
rotates straight in direction of 90 Degrees to the collision accordingly. Then display the alert on the
front of capturing walking direction of the screen to see whether it is very dangerous whether if
individual) attached to the individual’s mobile phone the collision happens. Some of the individuals were
along with the Default Camera of the device, but unable to sense if people are going into their
the image processing work is done on a different direction along or coming towards them from the
device because the main app with Alerts and the opposite direction. Some of them were unable to
other with Image Processing will put a lot of sense if the object was moving or static at one place.
pressure on the device’s processors. So, we can use a Individuals set their cameras manually, didn’t relied
Smart Watch as well and make its connectivity on automatic system and it worked as if it is in the
with the device through the Bluetooth so that the forward position then it coves almost the whole
Alerts App accessing the camera send images (per forward visualization that it was beeping right on
frame) to the Image Processing App installed on the time.
Smart Watch which an individual must wear or keep
close to the mobile device. 4.1 Future Workings
The CrashAlerts concept can even further go ahead
3.4 Task
as technology is moving fast and we can even be
So in order to carry out the final experiment to check able to implement Artificial Intelligence to
how beneficial and helpful this is going to be for the accomplish the task of alerting the individual if
individuals. We first took a survey from students he/she is about to collide to something. There is a lot
who are very much interacted with the mobile of room for this research to go further on in future
phones, in which we defined if having an application with even the sensors technology that might reduce
as such is helpful with some extra equipment such as the need of extra equipment needed in order to make
movable camera and smartwatch that is used that it work successfully.
they are fine with. Then we also asked some
students in the university to walk in the corridors 5. Conclusion
while consistently using their smartphones for any
purpose such as playing a game on the go, texting So we successfully created a Crash Alerts App with
someone, reading something etc. It was up to them better concept that helped reduce the heavy carriage
and that they shall act naturally by walking in the of equipment as an requirement for image
corridors in random directions to each other. There processing and capturing the direction. It helped the
were 8 participants in total who wore the individual to be alert before crashing into the objects
smartwatch, connected movable camera to their ahead while his/her eyes are busy with the
devices and started to walk. They even configured interaction to their mobile devices. This concept has
both the applications in both the devices been beneficial for the young generation as we
accordingly. figured out by doing the observation and then
choosing the student level individuals for the task to
perform and get the feedback about if it works. Even Practical Pedestrian Distraction Detection
we are definitely aware that it requires lot of further Framework using Wearables
improvements in future but has been successful till
4. This Causes 11,000 Injuries a Year—and You’re
now.
Probably Doing it Every Day
https://www.rd.com/culture/texting-and-walking-
References injuries/ last accessed 29th December 2018.
1. Mark B. Neider ∗ , Jason S. McCarley, James A.
5. Texting While Walking https://www.nytimes.com/
Crowell, Henry Kaczmarski, Arthur F. Kramer
2017/10/25/learning/texting-while-walking.html last
(2009) Pedestrians, vehicles, and cell phones
accessed 29th December 2018
2. Juan David Hincapié-Ramos Pourang Irani (2013)
6. Mumtazimah Mohamad , Md Yazid Mohd
CrashAlert: Enhancing Peripheral Alertness for
Saman , Muhammad Suzuri Hitam (2015) A Review
Eyes-Busy Mobile Interaction while Walking
on OpenCV
3. Nisha Vinayaga-Sureshkanth, Anindya Maiti,
Murtuza Jadliwala, Heena Rathore (2017) Towards a

You might also like