Abu Dhabi University

Capstone Project

Progress Report
Hand Gesture Controlled Emergency Aerial
Assistance Using Smartphone Based 4G
Quadcopter

Supervisor:
Team Members:
Dr. Mohammed Assad Ghazal
Muhammad Obaidullah
Acting Chair, Department of
1030313
Electrical and Computer
Sifat Sultan 1003289
Engineering
B.Sc. Electrical Engineering
Lab Instructors:
Abu Dhabi University, U.A.E.
Eng. Ahmed Sweleh
Communication Major
Eng. Ibrahim Ibrahim

May 25, 2014

In The Name of Allah, The Most Beneficent, The Most Merciful.

1
Contents
1 Project Background 4

2 Problem Discussion 5

3 Project Description 5
3.1 Using a UAV for agility and accessibility . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.2 Use of a quadcopter for ease of control and high payload delivery . . . . . . . . . . . 7
3.3 Global control by using on-board 4G smartphone . . . . . . . . . . . . . . . . . . . . 8
3.4 Live streaming video vision by using UDP . . . . . . . . . . . . . . . . . . . . . . . . 8
3.5 Smartphone to Controller Board USB Connection . . . . . . . . . . . . . . . . . . . 9

4 Software Development Progress 9

4.1 Smartphone to AtMega 328 communication protocol development . . . . . . . . . . . 9
4.2 Leap Motion frame decoding and GUI binding . . . . . . . . . . . . . . . . . . . . . 11
4.3 Google Map for Quadcopter Navigation . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.3.1 Getting the SHA1 Fingerprint For Google Map . . . . . . . . . . . . . . . . . 15

5 Hardware Development Progress 17

5.1 3D Model Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
5.2 Gyroscope Sensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5.3 An unexpected tragedy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5.3.1 Temperature Sensitivity of PLA plastic used in 3D printer . . . . . . . . . . . 19
5.3.2 Strength of PLA plastic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5.4 New Body Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
5.5 Circuit Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
5.5.1 Circuit Design Version 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
5.5.2 Circuit Design Version 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
5.6 PCB Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
5.7 Hand Gesture Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
5.8 Kalman Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

6 Testing Results 26
6.1 Custom 3D Designed Body . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
6.1.1 Version 1 - PLA Plastic, Unchanged print . . . . . . . . . . . . . . . . . . . . 26
6.1.2 Version 2 - PLA Plastic, Modified print . . . . . . . . . . . . . . . . . . . . . 26
6.1.3 Version 3 - Carbon Fiber Body . . . . . . . . . . . . . . . . . . . . . . . . . . 26
6.2 Leap Motion Library for Gesture Control . . . . . . . . . . . . . . . . . . . . . . . . 26
6.3 Payload Capability calculation using MATLAB . . . . . . . . . . . . . . . . . . . . . 27
6.4 Weight Calculation of the Quadcopter. . . . . . . . . . . . . . . . . . . . . . . . . . . 28

7 Updated Cost Table 28

8 Gantt Chart 29

9 Tasks 31

10 Conclusion 38

2
 Abstract
This project aims to solve the problem of providing emergency medical or any other type of
assistance in a un-accessible area by using a 4G quadcopter which is controlled by hand gestures.
The hand gestures provided by the user will be used to maneuver the quadcopter to reach the
emergency location and provide assistance. The uniqueness of this project lies in the fact that
it uses a 4G smartphone to connect the quadcopter to the base station for getting the user
commands, providing the GPS location, and live video stream of the quadcopters vision.

1 Project Background
The number of researches and developments on the unmanned aerial vehicles have increased over
the past years. Many hobbyists and robotics enthusiasts are particularly interested in quadcopters
because of their ability to hover in place, take-off and land vertically. [4] People can easily now buy
and assemble a quadcopter with low cost. Within seconds a quadcopter can lift-off reach a point
and land back. This ultra fast response and ease of control make the quadcopter ideal for reaching
areas dangerous for humans and performing tasks.

Parallel to this, the field of Human-Computer Interaction has been a growing field in the past
few years due to the introduction of innovative interaction devices such as Nintendo Wii Remote
and Microsoft Kinect. The idea of using gestures as a mean of interaction has become the recent
trend as it makes the experience more natural and hence the possibility of application.

Quadcopter: The term quadcopter or quadrotor comes from the four narrow chord propellers
used by the air vehicle in order to maneuver in air. The quadcopter uses these four propellers to
push the air downwards in order to generate lift force. It maneuvers by varying the speeds of these
four motors.

Leap Motion: In 2012, the Leap Motion controller was introduced. Leap motion comes with a
150 degree field of view and makes use of a depth sensor to follow and scan the hand features up
to 1/100th of millimeter. This extremely precise control gives us the opportunity to use it as our
interface over Kinect which tracks body instead of hand since our gestures will be mainly finger
movements.

Figure 1: A leap motion device with 3D coordinates.

The past few years there have been extensive research on autonomous micro-aerial vehicles. The
application ranges from hobby to serious circumstances.

3
Quad Copter Application XPROHELI developed quad copter that is used to take high quality
video and photography. http://www.youtube.com/watch?v=HoQQlRzybmA#t=14 This video was
taken using the XPROHELI quad copter.

FAE developed tri copter that is used to take long range aerial surveillance over oil field. http:
//www.youtube.com/watch?v=a5H5paygWQw&feature=youtu.be Quadcopter has the potential to
be used in a variety of application and we decided to work on a case where quad copter deserves
to be used; to save human lives where time is of paramount.

2 Problem Discussion
In this stone hearted world, everyday accident happen and people die. These lives can be saved
by timely medical response. But the response faces countless obstacles sometimes like traffic jams,
remote places, radiation etc. This is a growing problem causing engineers and specialists to think
and come up with different solutions.

UAE itself has a mortality rate of about 37 per 100,000 which is one of the highest in the world.
[1] This mortality rate can be further skimmed down by providing timely first aid kit to seriously
injured people.

Figure 2: Lives are lost as fast-aids do not reach on time.

So all in all, our problem statement is how to provide quick and timely emergency assistance
without the restriction of the range by using low-cost quadcopter.

3 Project Description
We proposed a system which is complete package and will solve all the problems of providing
quick emergency assistance. The proposed system was widely accepted by the faculty members
and approved.

4
 Figure 3: The system diagram of the project.

Our system has a Quadcopter with a gyro control chip which controls the basic maneuvers for
the UAV and we propose to build a custom designed chip to decode the signals coming from
the on-board android smartphone and send commands to the controller chip. We used a android
smartphone to provide the quadcopter with GPS, Camera, and 4G internet connection. So this
quadcopter is now capable of performing completely unmanned flight and be controlled using http
requests coming to it through 4G connection. The quadcopter also provides a http feedback to
the computer about its GPS position. The GPS can also be used to make a flight plan for the
quadcopter. The on-board mobile also provides a UDP live stream of the camera vision and can
be controlled using.

Lets take a typical emergency assistance scenario and apply our proposed solution to it. A call
comes to the police about a car accident that has happened on a highway in Abu Dhabi. This
is emergency case it might take several hours for the police to arrive and analyze the situation.
The police places a pin on the map in the quadcopter software in computer. The computer sends
the http request to the quadcopter to move to that specific location. The quadcopter immediately
starts up and goes to the GPS location. This is done with a feedback loop which calculates the
current GPS location and the required GPS location. While all of this is done, the policeman can
see the current location of the quadcopter on the map. When the quadcopter reaches the end of its
destination, the policeman turns on the gesture control mode and starts controlling the movement
of the quadcopter precisely using hand. Now the policeman can access the situation and can also
possibly drop the first aid kit if the situation is too serious.

3.1 Using a UAV for agility and accessibility

Unmanned Air Vehicles are light, fast and can be designed to carry the essential aid equipment.
Along with that unmanned vehicles can be used to perform critical operations in areas where human
life cannot be risked. Furthermore, quick response to a distress call can differentiate between life
and death of a person. Therefore, air emergency medical assistance is needed very quickly in some
cases and the delay cased by set-up and start of manned vehicles cannot be afforded at this point.

5
3.2 Use of a quadcopter for ease of control and high payload delivery
With the use of four propellers pushing the air downwards with immense speed, a typical quad-
copter is capable of lifting high payloads. UAVs for search and rescue operations are required to
be small and at the same time carry high payloads. Quadcopter is perfect for such kinds of ease
of control applications because of its usage of four high rpm brushless motor run propellers. These
brushless motor speeds can be intelligently varied to provide 6 degrees of freedom by sensing the
gyroscopic values of Pitch, Yaw, and roll which.

The quadcopter can be controlled using two types of configurations, one is the X configuration
and the other is the + configuration. For a very stable flight and camera view being not blocked,
we propose to use a X configuration quadcopter.

Figure 4: X configuration forward motion of Figure 5: Left motion of quadcopter in X con-

quadcopter. figuration

Forward Direction: By speeding up the back motors, a forward direction motion can be ac-
complished. This makes the thrust of the motors at an angle to the z-axis. The thrust has force
components in the weight direction to balance or cancel the weight and a unbalanced forward
component to make the quadcopter go forward.

Left Direction: By speeding up the right motors, a left direction motion can be accomplished.
The thrust has force components in the weight direction to balance or cancel the weight and a
unbalanced forward component to make the quadcopter go left.

Right Direction: By speeding up the left motors, a right direction motion can be accomplished.
The thrust has force components in the weight direction to balance or cancel the weight and a
unbalanced forward component to make the quadcopter go right.

6
Figure 6: X configuration forward motion of Figure 7: Left motion of quadcopter in X con-
quadcopter. figuration

Backward Direction: By speeding up the front motors, a backward direction motion can be
accomplished. This makes the thrust of the motors at an angle to the z-axis. The thrust has force
components in the weight direction to balance or cancel the weight and a unbalanced backward
component to make the quadcopter go forward.

3.3 Global control by using on-board 4G smartphone

Usually quadcopters have a limited range radio link but emergency assistance should not have
limited range of control. It was decided to use 4G network for control and communication with the
quadcopter as 4G network is accessible from within 35Km range of a cellular tower. Furthermore,
a typical city is divided into cells and at almost each node a telecommunication tower is located.
This is a huge advantage for our quadcopter since it can have a connection to base control all
around the city where the 3G radiation pattern is in range. A smartphone will be mounted on the
quadcopter for providing a 4G internet connection. The internet connection on the quadcopter can
be used for several things such as providing the GPS location, video stream, and controlling the
quadcopter maneuvers.

3.4 Live streaming video vision by using UDP

Video streaming of quadcopters vision is possible by using the on-board android smartphone.
An app it to be designed in such a way that it streams video back to the base station by use of UDP.
This is possible through 4G connectivity since use of WiFi for such long range is not advisable.

A video stream is acceptable if there is any fault of packages lost in the way. It is because if
during the video streaming the packages get lost due to the router being overloaded for instance
then the video streaming can still be resumed with a reduced quality.

7
 The TCP/IP protocol would force the video stream to wait until the dropped packages are found
to resume processing newer packages. The point that is usually misunderstood is that when few
packages are lost it doesnt mean a whole frame is lost but rather few pixels from a certain frame
are lost and hence waiting for these lost pixels to come back is a waste since the video stream has
already moved on to new frames. However TCP could have been a good option if the case was
about recorded video streaming since when a package is lost the system pauses the frame until the
all the data for the next frame has been collected in the buffer giving the term buffering.

Therefore, we decided to use UDP for our live video streaming.

3.5 Smartphone to Controller Board USB Connection

The receiver in our schematics is going to be a smartphone which will be using 4G internet to
receive the data. The smartphone onboard the quadcopter will act as client while the PC with the
user will act as the server to send the controls. The data will then be passed to an Atmega328
chip using USB-Serial FTDI chip that will in turn send the flight data to controller board. The
application will send byte array through serial to Atmega328 Chip that will contain the flight
values such as aileron value to control the roll, elevation value to control the elevation angle with
horizontal, rudder value to control the yaw movement, and thrust value to control lift force. [?]

Figure 8: A custom designed and programmed AtMega 328 chip will act as a intermediate connector
to decode the signals coming from the android smartphone through USB and send appropriate
commands to the control board.

4 Software Development Progress

4.1 Smartphone to AtMega 328 communication protocol development
Since our approach of controlling the quadcopter using the smartphone is new and little or no
documentation is present on how to send the data through Android to quadcopter, we had to
develop our own network communications protocol. The figure below gives and overview of the
OBFlightPacket network protocol we developed.

8
Figure 9: The figure shows how an AtMega 328 is connected to the FTDI USB-Serial converter
chip. [2]

Figure 10: The figure explains the network protocol which was designed by us for sending flight
data from Android USB to micro-controller.

OBFlightPacket Network Protocol Once the communication link is established, the android
starts sending the OBFlightPacket. The OBFlightPacket is nothing but a byte array of size 7.
As the byte is composed of 8 bits, there are 28 = 256 possibilities. So if we convert from byte
to unsigned integer, the value can vary from 0 - 255. If we convert the byte to ASCII character
(American Standard Code for Information Interchange), each of these possibilities (symbols) can
represent all charecters on keyboard and some special characters too.

Now the micro-controller AtMega 328 can communicate with the smartphone using the USB
connection. The control data can be transferred back and forth between the smartphone and
AtMega. There is a library in Android to communicate with the smartphones USB port called
USBManager.

We used this library in order to send the byte array to the USB port. A USB-serial converter
chip called FTDI is connected to the smartphones USB port and the FTDI chip is then connected

9
 Figure 11: The figure which was developed in Photoshop explains the algorithm to unpack the
OBFlightPacket. There are several checking stages and even if one checking results in false, the
packet is rejected.

to XBee.

Figure 12: The figure shows one of the ways the smartphone can communicate with the AtMega328.
The XBee module is attached using a USB cable to the smartphone and the XBee explorer which
has FTDI chip on it, takes care of the conversion of data signals from USB to Serial.

4.2 Leap Motion frame decoding and GUI binding

A leap motion device was bought and the API for grabbing frame data was used to develop a
user-friendly GUI for visualizing the Pitch, Yaw, Roll angles. In the Application Programming
Interface (API), the leap motion provides vectors for each finger positions and access to several
other objects such as hands, fingers, palm normal, palm direction, and palm Position.
1 p r i v a t e v o i d LeapFrameUpdate Tick ( o b j e c t s e n d e r , EventArgs e )
{
3 // Get t h e c u r r e n t frame o b j e c t from t h e Leap Motion C l a s s
currentHandFrame = c o n t r o l l e r . Frame ( ) ;

10
 5 // E x t r a c t i n g v a l u a b l e i n f o r m a t i o n from fram and d i s p l a y i n g on TextBoxes
frameIdBox . Text = currentHandFrame . Id . T o S t r i n g ( ) ;
7 timeStampBox . Text = currentHandFrame . Timestamp . T o S t r i n g ( ) ;
handsBox . Text = currentHandFrame . Hands . Count . T o S t r i n g ( ) ;
9 f i n g e r s B o x . Text = currentHandFrame . F i n g e r s . Count . T o S t r i n g ( ) ;
g u e s t u r e B o x . Text = currentHandFrame . G e s t u r e s ( ) . Count . T o S t r i n g ( ) ;
11
// Do t h i s i s t h e c u r r e n t frame c o n t a i n s hands
13 i f ( ! currentHandFrame . Hands . IsEmpty )
{
15 // Get t h e f i r s t hand
Hand hand = currentHandFrame . Hands [ 0 ] ;
17
// Check i f t h e hand has any f i n g e r s
19 F i n g e r L i s t f i n g e r s = hand . F i n g e r s ;
i f ( ! f i n g e r s . IsEmpty )
21 {
// C a l c u l a t e t h e hand s a v e r a g e f i n g e r t i p p o s i t i o n
23 Leap . Vector avgPos = Leap . Vector . Zero ;
foreach ( Finger f i n g e r in f i n g e r s )
25 {
avgPos += f i n g e r . T i p P o s i t i o n ;
27 }
avgPos /= f i n g e r s . Count ;
29 }

31 // Show t h e 3D c o o r d i n a t e p o s i t i o n o f hand i n t e x t box

P o s i t i o n B o x . Text = hand . P a l m P o s i t i o n . T o S t r i n g ( ) ;
33
// Get t h e hand h e i g h t from s e n s o r and check i t s v a l i d r a n g e
35 i f ( ( hand . P a l m P o s i t i o n . y < 5 0 0 ) && ( hand . P a l m P o s i t i o n . y > 0 ) )
{
37 Thrust = ( i n t ) ( ( ( hand . P a l m P o s i t i o n . y / 5 0 0 . 0 ) 7 1 . 0 ) + 5 7 . 0 ) ;
}
39
// Get t h e hand h e i g h t from s e n s o r and show i n p r o g r e s s bar
41 i f ( ( hand . P a l m P o s i t i o n . y < 7 0 0 ) && ( hand . P a l m P o s i t i o n . y > 0 ) )
{
43 p o s i t i o n P r o g r e s s . Value = ( i n t ) hand . P a l m P o s i t i o n . y ;
}
45
// Get t h e hand s normal v e c t o r and d i r e c t i o n
47 Leap . Vector normal = hand . PalmNormal ;
Leap . Vector d i r e c t i o n = hand . D i r e c t i o n ;
49
// D e f i n i n g a c e n t e r p o i n t f o r t h e arrow p i c t u r e
51 System . Drawing . P o i n t p = new System . Drawing . P o i n t ( ) ;
p .X = 1 2 8 ;
53 p .Y = 1 2 8 ;
System . R e f l e c t i o n . Assembly t h i s E x e ;
55 t h i s E x e = System . R e f l e c t i o n . Assembly . GetExecutingAssembly ( ) ;
// Loading t h e arrow png image f o r n i c e a n g l e r e p r e s e n t a t i o n
57 System . IO . Stream f i l e = t h i s E x e . G e t M a n i f e s t R e s o u r c e S t r e a m (
O r e g o C o n t r o l l e r . P r o p e r t i e s . R e s o u r c e s . a r r . png ) ;

59 f l o a t p i t c h V a l = ( d i r e c t i o n . P i t c h 1 8 0 . 0 f / ( f l o a t ) Math . PI ) ;
// Rotate t h e arrow a c c o r d i n g t o t h e p i t c h v a l u e
61 p i c t u r e B o x 2 . BackgroundImage = RotateImage ( g l o b a l : : O r e g o C o n t r o l l e r .
P r o p e r t i e s . Resources . arr , p , pitchVal ) ;

11
 // D i s p l a y p i t c h i n t e x t b o x
63 pitchBox . Text = p i t c h V a l . T o S t r i n g ( ) ;
// Check t h e p i t c h v a l u e r a n g e and s e t i t
65 i f ( ( p i t c h V a l < 9 0 ) && ( p i t c h V a l > 90) )
{
67 E l e v a t i o n = ( i n t ) p i t c h V a l +90;
}
69
f l o a t r o l l V a l = ( normal . R o l l 1 8 0 . 0 f / ( f l o a t ) Math . PI ) ;
71 // Rotate t h e arrow a c c o r d i n g t o t h e r o l l v a l u e
p i c t u r e B o x 3 . BackgroundImage = RotateImage ( g l o b a l : : O r e g o C o n t r o l l e r .
P r o p e r t i e s . Resources . arr , p , r o l l V a l ) ;
73 // D i s p l a y r o l l i n t e x t b o x
r o l l B o x . Text = r o l l V a l . T o S t r i n g ( ) ;
75 // Check t h e r o l l v a l u e r a n g e and s e t i t
i f ( ( r o l l V a l < 9 0 ) && ( r o l l V a l > 90) )
77 {
A i l e r o n = ( i n t ) r o l l V a l +90;
79 }

81 f l o a t yawVal = ( d i r e c t i o n . Yaw 1 8 0 . 0 f / ( f l o a t ) Math . PI ) ;

// Rotate t h e arrow a c c o r d i n g t o t h e yaw v a l u e
83 p i c t u r e B o x 4 . BackgroundImage = RotateImage ( g l o b a l : : O r e g o C o n t r o l l e r .
P r o p e r t i e s . R e s o u r c e s . a r r , p , yawVal ) ;
// D i s p l a y yaw i n t e x t b o x
85 yawBox . Text = yawVal . T o S t r i n g ( ) ;
// Check t h e yaw v a l u e r a n g e and s e t i t
87 i f ( ( yawVal < 9 0 ) && ( yawVal > 90) )
{
89 Rudder = ( i n t ) yawVal ;
}
91 }
}

In the ground base station, following things can be done. In other words, the features of the
ground station are as following:

Frame is being captured from the leap motion device.

After extensive image processing by leap motion library on the current frame, the library
initializes hands, fingers, and palms objects in the frame parent object and returns frame for
programmer to access and use.

This frame object is accessed by the C# application and some of the data enclosed is displayed
on the screen in forms of text-boxes, labels and pictures.

The arrow picture is loaded and rotated to the same angle according to the Pitch of one
the hands detected by the leap motion.

The arrow picture is loaded and rotated to the same angle according to the Roll of one the
hands detected by the leap motion.

The arrow picture is loaded and rotated to the same angle according to the Yaw of one the
hands detected by the leap motion.

12
 The hand y position in 3D is extracted from the current frame and displayed in a textbox
and a progress bar indicates how high above is the hand approximately from the sensor. The
maximum sensitivity we detected using our Leap Motion device was about 70cm.

The current frame time-stamp, frame ID., number of hands detected, number of fingers
detected, and gesture detected state are extracted from the frame object and displayed using
the text-boxes.

Pitch, Yaw, Rudder, and Thrust values are then converted to byte array according to the
Flight protocol which we discussed earlier.

The GPS way points for a quadcopter can be set and stored in the database.

Figure 13: The figure explains a groupbox which is embedded inside the main application and
contains the GUI for viewing the Leap Motion data and variables.

13
Figure 14: The overall C# application main window with all the settings and options. The ap-
plication allows communication through internet and XBee as well. The commands can be sent
manually by entering it into textboxes or directly decoded from hand gestures.

4.3 Google Map for Quadcopter Navigation

4.3.1 Getting the SHA1 Fingerprint For Google Map
To work with Google map we need to use the API; however; since the Google map API is not
intended to be open source and only belong to Google in order to be changed the API comes as a
part of Google Play Services. It simply means we need permission from the Google Play Service to
allow us use their Google Map API.

The process of obtaining the permission is by first letting my app to contact with the Goolge
API site. This is made possible by the following steps:

14
Figure 15: We generate an APK that will have a SHA1 finger-print. This finger-print will be stored
in a .jks file format which cannot be opened except with the keytool.exe application that comes
along when one downloads the Java SDK.

In order to tell keytool to open that .jks file we use cmd as the interface. Below are a series of
simple commands that we need to type in order to open keytool application and then tell it to open
my .jks file to retrieve the SHA1 finger print

Figure 16: Signed APK was generated and stored the key in the following file Now it was time to
retrieve the key that is stored inside the googlemap.jks

Step 1: We entered the path of keytool by typing the following commands

cd C : \P rogramF iles\Java\jdk1.7.0 51\bin
Step 2: Then we retrieved the SHA1 Fingerprint in the following way:

15
5 Hardware Development Progress
5.1 3D Model Design
A 3D Model of the quadcopter was custom designed to fit our needs. The 3D Model was supposed
to have arms of specific lengths to allow the propellers to rotate freely. We designed the whole
quadcopter in Google Sketch Up taking care of these parameters. The design also was supposed to
have a holder for the mobile phone. We designed a intelligent mechanism for controlling the tilt
angle of the mobile. We used 10 inch diameter propellers with 3.8 inch of pitch. The pitch controls
the amount of air scoped with each rotation. The diameter controls two things, amount of air the
propeller pushes down, and the weight of the propeller. Careful calculations and testings were done
to calculate the amount of weight of each motor can lift.

Figure 17: A arm of the quadcopter is being printed by the MakerBot.

Features of the Model:

Four less infill light 10 motor arms.
Three center platforms with screw holes to mount the control chip.

16
 A raft to hold and rotate the mounted android smart-phone.
Four stands for landing gear.

5.2 Gyroscope Sensor

A quadcopter is highly prone to instability due to irregular distribution of mass or wind.Basic
moves of a quadcopter like lifting straight off the ground to complex moves like pitch(right) and
roll or yaw cannot be achieved without stabilizing the quadcopter.

Figure 18: Explanation of the traditional 3 Dimensional rotation movements. These movements
were defined for airplanes but as time passed by, these same terminologies were adopter by quad-
copters although quadcopters dont bend their wings for roll(aileron).

It has become the industry standard to implement sensors to know the precise measurment of
imbalance instantly and exert the proper combination of force to bring it back to balance. These
sensors are known as Gyroscope and they use the law of physics to measure this phenomenon.

Figure 19: The gyroscope sensor is designed to measure the rotation per second and the angle of
rotation of the object its mounted to when it experiences an imbalance.Below is a typical 3-axi
gyroscope ITG3200. [6]

17
 Hence, when the object tilts a little to the right(roll), the gyroscope on the y-axis will record a
reading while the other three will not record any reading, or when the device tilts up(pitch) the
sensor in x-axis will record a value while the rest wont record any value and finally if the device
rotate in the horizontal plane the sensor in z plane will measure a current due to angular rotation
around the z-axis

The working of the gyroscope is as fascinating as it is clever. There is a small mass inside each
gyroscope on a particular axis which is held by very small springs. When there is a rotation, this
will cause a centripetal force to form whose direction will depend on the direction of rotation. For
instance; if the gyroscope is rotating right there will be a force that is exerted toward the center
while it the gyroscope is rotating outward there will be a force that opposite to the center.

Figure 20: These forces are exerts the spring and this movement generates very low-current electrical
signal that is amplified and later read by the micro-controller. [6]

5.3 An unexpected tragedy

After designing the whole body using the labs 3D printer, the quadcopters body melted and
twisted away one day due to sun heat. This gave us a strong lesson that the plastic used to manu-
facture the quadcopter body is very sensitive to the high temperatures. Since UAEs temperature
is usually very high in major part of the year, the 3D plastic cannot be used for building the
quadcopter.

5.3.1 Temperature Sensitivity of PLA plastic used in 3D printer

The plastic which is used in the 3D printer is Polylactic Acid thermoplastic aliphatic polyester.
It is usually derived from renewable resources such as corn starch. It has a melting point of 150-
160 degrees centigrade. The temperature in UAE reaches 45 degree centigrade in summers and
even with a small amount of insulation, this temperature can reach high enough to just twist or
deform the plastic. The choice of PLA plastic for quadcopter body is really bad when it comes to
temperature sensitivity.

5.3.2 Strength of PLA plastic

In the first two-three attempts, the quadcopter crashed several times and broke the arms within
milliseconds after coming in contact with the ground. This was also a big issue as we want the final
product to have strength enough to withstand huge jerks and shocks in case the quadcopter falls
down from huge height.

18
Figure 21: The table shows strength of the plastic used for 3D printing in our lab according to
different standards. [3]

5.4 New Body Design

After getting the lesson, we changed the body from PLA 3D printed plastic to Carbon Fiber body.
The body was bought as a kit from HobbyKing named Turnigy Talon Carbon Fiber Quadcopter
Frame. Link: http://hobbyking.com/hobbyking/store/__22397__Turnigy_Talon_Carbon_Fiber_
Quadcopter_Frame.html

5.5 Circuit Design

5.5.1 Circuit Design Version 1
Initially when we started our research on Quadcopter controller board, we found that we needed a
PCB to find out what signals exactly are required by the quadcopter in order for it work perfectly.
Additionally, connecting and disconnecting all the wires on the breadboard was very hectic and
time consuming. For this purpose we created the version 1 of the PCB. It was the pre-final version
of the PCB. This PCBs job is to transmit the appropriate Aileron, Rudder, Thrust and Elevation
PWM signals to the controller board so that it can further perform PID control on brush-less
motors. Essentially, this PCB is a bridge between the controller board and the smartphone. The
following were the features of the first circuit.

Servo connectors for four ESCs.

Servo connector for mini servo which will be used for changing the mobile angle.

Reset push button.

Tilt Compensated Compass output pins.

SPI programming header.

16 MHz crystal with 22pF capacitors for providing clock signal to AtMega328.

19
 Ultrasonic height sensor connection.

FTDI connection pin.

3.3V Regulator for stepping down he voltage from 5V to 3.3V for the XBee.

3 indicator LEDs of different colors.

XBee connection support.

Bluetooth module connection support using the FTDI pins.

Figure 22: A circuit diagram of our prototype board design version 1.

5.5.2 Circuit Design Version 2

We quickly realized that several improvements can be done to the initial circuit to make the
quadcopter have more and more features. The following are the additional features of the new
PCB.

IR Receiver for receiving Infra red signals. Basically this includes TV, CD player and other
IR transmitters. The IR receiver already has a electronic circuit to amplify the IR signal
and provide a signal when a change in IR level is detected. Also the Sensor already has the
background noise removal circuit. SO we can just connect the infra-red sensor directly to the
interrupt pin to read the bits transmitted and change the states of the quadcopter. This IR
Receiver will be used

5V Barrel Jack connector for providing 5V to the circuit.

USB A jack for supplying power.

USB B jack for supplying power.

20
 Figure 23: A circuit diagram of our prototype board design version 2.

5.6 PCB Design

The PCB was designed to be compact and at the same time to contain all the essential components
for communication. There are two ways a communication link can be made with the on-board
ATmega 328 on our PCB. The first method is by using XBee and the second method is by using
the USB Port.

Features of the Prototype PCB

4 servo connectors for connecting to the Electronic Speed Controllers.

8 Pin connector for connecting a tilt compensated compass.

6 pin connector for FTDI chip to convert USB data signals to serial data.

4 pin connector for connecting a height measurement sensor.

A 3 pin connector to connect the mobile tilting sensor.

A SPI JTAG programming header.

XBee connector for wireless communication.

3 LEDs to indicate the status of the AtMega microcontroller.

In the above figure, the prototype board is shown where we have used XBee to communicate
with the computer and receive the serial data to control the motor speed. This is just for testing
purposes and in the final design we will use the data coming from the android to determine the
speed of the motors. Furthermore, the final design will use the XBee as a kill switch to completely
turned off the quadcopter in case it reaches a forbidden area or any similar situation.

21
Figure 24: A x-ray vision of our prototype board design version 1.

Figure 25: A x-ray vision of our prototype board design version 2.

22
 Figure 26: A x-ray vision of our prototype board design version 3.

Figure 27: A x-ray vision of our final board design version 4.

5.7 Hand Gesture Models

In order to control the quadcopter using a leap motion device, a model of the hand gestures
should be made in order to approximate the finger positions in three dimensional space.

Leap Motion provides a library to use for creating hand objects and approximates the finger
positions in the 3 dimensional space. According to these positions of the fingers, we can code the
Leap Motion to generate http requests to the android on the quadcopter through internet.

23
 We will program a java application to detect the hand gestures and then according to the gesture
input, a particular http request will be made to the android. Eclipse IDE will be used to program
the desktop application which will provide the internet connection to the quadcopter.

Figure 28: Moving the hand upwards in the z- Figure 29: Moving the hand downwards in the
axis in front of the Leap motion should provide z-axis in front of the Leap motion should pro-
us with a +ve change in z value. vide us with a -ve change in z value.

Figure 31: A +ve value in x-axis can be ob-

Figure 30: Moving the hand towards the body tained when the user moves the hand to the
can give us -ve value of movement in y-axis. right.

Figure 32: A -ve value in x-axis can be ob- Figure 33: Moving the hand away from the
tained when the user moves the hand to the body can give us +ve value of movement in
left. y-axis.

5.8 Kalman Filter

There is a module in the controller board that processes the huge amount of samples the micro-
processors take from the sensors. To put that into perspective lets say one have a micro-controller
that has a clock speed of 15 MHz, that means the micro-controller will be taking 15 million sam-
ples from the sensor every second. All this amount of data is meaningless if the right data is not
processed from this, Kalman Filter is a sequence of steps that will initially remove the sampled
data that are purely noise and then average out the important data to pick the most appropriate
one for input.

24
 The leap motion captures about 200 frames per second or more depending upon the computer
processing power. This is huge data speed and sudden errors can cause the system to In our project,
we will implement a Kalman Filter on the hand gesture data.

6 Testing Results
6.1 Custom 3D Designed Body
Successful printing of the body was achieved but arms had to be manually drilled for the holes
so that the motors fit precisely. The body was assembled using the screws with washers to prevent
the plastic to breaking down. The hols for the circuit board to mount were made and the place for
the mobile angle mover servo was also cut out.

In the website from Makerbot named Thingiverse (www.thingiverse.com), there are several 3D
models designed for several sized quadcopters. Although these models are similar to what our
final product required, these models need to be modified in several places to fit out needs and
requirements.

6.1.1 Version 1 - PLA Plastic, Unchanged print

We printed the 3D quadcopter model as is from the thingiverse website without modifying any
part of the quadcopter. This was done in order to make sure that the existing printer in the lab
can print the model, all the parts fit together, and allow additional electronic components to be
mounted onto the central body.

We found that all the parts indeed fit together but a major increase in the height of the arms
from ground was needed. Also the motors we had did not fit on top of the arms perfectly. An
increase in the diameter of motor mounts was also needed.

6.1.2 Version 2 - PLA Plastic, Modified print

We increased the arm motor mount diameters and increased the height of the landing pads and
printed the parts using the 3D printer.

We found that all the parts indeed fit together but a tragedy occured when we left the model in
the car for two days. Under the heat of the sun and no air flow, the quadcopter body melted and
got twisted.

6.1.3 Version 3 - Carbon Fiber Body

We bought the carbon fiber body from Hobby King and made sure that the dimensions matched
rest of the parts. The carbon fiber is composite material which is many times stronger than steel
and quite light in weight.

6.2 Leap Motion Library for Gesture Control

There is a huge documentation online for programming the leap motion controller in Java. The
gesture detection can be achieved using a Controller Object which can return the hand tracking
data by using frame object within this class. [5] For example in the following code:

25
 S a m p l e L i s t e n e r l i s t e n e r = new S a m p l e L i s t e n e r ( ) ;
2 C o n t r o l l e r c o n t r o l l e r = new C o n t r o l l e r ( ) ;
// Have t h e sample l i s t e n e r r e c e i v e e v e n t s from t h e c o n t r o l l e r
4 c o n t r o l l e r . addListener ( l i s t e n e r ) ;
Frame frame = c o n t r o l l e r . frame ( ) ;
6 System . out . p r i n t l n ( Frame i d : + frame . i d ( )
+ , timestamp : + frame . timestamp ( )
8 + , hands : + frame . hands ( ) . count ( )
+ , f i n g e r s : + frame . f i n g e r s ( ) . count ( )
10 + , t o o l s : + frame . t o o l s ( ) . count ( ) ) ;

6.3 Payload Capability calculation using MATLAB

A MATLAB code was written according to the propeller efficiency, power of the motors and the
propeller size to calculate the payload the quadcopter is capable of carrying.
% P r o p e l l e r hover e f f i c i e n c y
2 eta = 0 . 7 5 ;
% Power o f t h e motor Max . f o r our motor i s 125
4 Power = 1 1 0 ;
% P r o p e l l e r Radius i n m e t e r s d i a m e t e r = 10 i n c h e s = 0 . 2 7 9 4
6 R = 0.2540;
% Usual Air D e s n i t y kg /m3
8 rho = 1 . 2 2 ;
Thrust = ( ( e t a+Power ) 2 2 p i R2 rho ) ( 1 / 3 ) ;
10 d i s p ( Thrust i n Newtons : ) ;
Thrust
12 d i s p ( Weight L i f t a b l e by one motor i n Kg : ) ;
Weight = Thrust / 9 . 8 0 6 6 5 0 0 2 8 6 4 ;
14 Weight
d i s p ( Weight L i f t a b l e by a l l f o u r motors i n Kg : ) ;
16 Weight = ( Thrust / 9 . 8 0 6 6 5 0 0 2 8 6 4 ) 4 ;
Weight

The following was the result of the simulation:

26
 Figure 34: Moving the hand away from the body can give us +ve value of movement in y-axis.

6.4 Weight Calculation of the Quadcopter.

A weight load estimation excluding the extra payload was tabulated:

Component Type Qty. Unit Total

ESC 30Amps 4 34.92g 139.7g
Motor & propellers Super Tigre 4 59.45g 237.8g
Plastic Frame 3D Custom 1 248g 248g
Screws Metallic 14 9.0g 125.0g
Servo Tiny 1 11.3g 11.3g
Control Brd. & wires Hobby King 1 21.0g 21.0g
AtMega Brd. Custom 1 29.9g 29.9g
Android Smartphone Huawei P6 1 120g 120g
11.V Battery ZOP 1 369.7g 369.7g
Ultrasonic Sensor N/A 1 10.2g 10.2g
Wiring N/A 1 20g 20g
Total 1332.6g
The weight calculation showed that the quadcopter parts were no heavier than 1332.6 grams.
in order for a quadcopter to have nice thrust, typically the thrust produced by the motors should
be twice the weight. Our first test runs of the motors show that the quadcopter is easily capable
of lifting these weights.

7 Updated Cost Table

27
Figure 35: The table shows all the parts which were bought from various sources and countries
including parts which were bought and brought personally from canada.

8 Gantt Chart

28
29
9 Tasks
Task Name: Modifying 3D body design
Accomplished By: Muhammad Obaidullah
Starting Date: 4 December 2013
Completion Date: 26 December 2013
Obstacles Faced:

1 The Sketch Up did not allow modification of the motor mount unless the landing
pad was removed. Used an older version of the Sketch Up file provided online and
printed the leg separately.

Percentage Completion: 100%

Output: The quadcopter arms and body is now completely designed in Google Sketch
Up and can be exported as .stl format for printing using the Makerbot.

Task Name: Printing the 3D model and painting it

Accomplished By: Muhammad Obaidullah
Starting Date: 30 December 2013
Completion Date: 26 March 2014
Obstacles Faced:

Several times, the electricity of the lab went and the model had to be printed again.

After repeated printing and experimenting, the size of the holes and arms was mod-
2 ified for perfect fitting.

Several test flights resulted in broken parts of the quadcopter, these parts were
printed again.

Percentage Completion: 100%

Output: PLA plastic body is completely printed using the labs 3D printer and painted
using spray paint.

30
 Task Name: Reverse engineering the HobbyKing flight control board
Accomplished By: Muhammad Obaidullah
Starting Date: 28 January 2014
Completion Date: 12 March 2014
Obstacles Faced:

Surface mount resistor values were very hard to find out because of small size.

3 The signals required by the flight control board were found and confirmed using the
oscilloscope.

Percentage Completion: 100%

Output: We know exactly what signals are coming out from the HobbyKing flight
control board and what input signals (Elevation, Aileron, Thrust, Rudder) signals are
required as inputs. It was also of utmost importance to find out the Vpp and duty cycle
of the output PWM which is fed into the ESCs (Electronic Speed Controllers).

Task Name: AtMega 328 to Android Protocol Development

Accomplished By: Muhammad Obaidullah
Starting Date: 12 March 2014
Completion Date: 26 March 2014
Obstacles Faced:

Serial communication gives low error rate under slow bit rate and XBee performs
4 well close to 1200 bps serial data rate. Therefore the baud rate was reduced to 1200
bps. And appropriate delay was added in the Serial data read loop to wait for the
next byte to arrive.

Percentage Completion: 100%

Output: Using this protocol, the quadcopter can be controlled completely. The android
smart-phone can send flight data and control the global positioning of the quadcopter.

31
 Task Name: Configuring XBee for serial communication
Accomplished By: Muhammad Obaidullah
Starting Date: 26 March 2014
Completion Date: 27 March 2014
Obstacles Faced:

New version of XCTU does not recognize the XBees bought. Downloaded older
version of XCTU and configured.
5 The baud rate was reduced to 1200 bps to ensure low bit error rate. This rate was
chosen after many tests using the oscilloscope channel 1 connected to the RX pin of
XBee to detect whether it received every byte sent at different data speeds.

Percentage Completion: 100%

Output: The quadcopter can now be controlled by using either XCTU, C# windows
application, or Android application. The following link was used: http://www.science.
smith.edu/~jcardell/Courses/EGR328/Readings/XbeeGettingStarted.pdf

Task Name: Capture video using Camera object

Accomplished By: : Sifat Sultan
Starting Date: 1 January 2014
Completion Date: 5 June 2014
Obstacles Faced:

Our quad copter is required to take a video stream of the surrounding and using
any USB camera that are found in the market to carry out this task is bad choice as
6 with every addition of weight the battery life of the quad copter is bound to reduce.

Percentage Completion: 80%

Output: Since the quad copter comes with smartphone which is intended to do the task
of providing with 4g communication service, it only makes sense to use the other features
that comes along with a standard android smartphone. It comes with a powerful built
in camera that can capture HD video. Therefore we programmed the app to open the
camera and capture the video through the following code.

32
 Task Name: Play Live Stream using HttpURLConnection
Accomplished By: : Sifat Sultan
Starting Date: 15 January 2014
Completion Date: 5 June 2014
Obstacles Faced:

The smartphone will stream the video to a certain IP and Port address. This stream
is very efficiently encoded such that it needs powerful library to be decoded.

I planned to use a WebView or a Media Player to play the video streaming. However;
7 neither WebView nor Media Player in android is not powerful enough to decode the
stream and play video.

Percentage Completion: 70%

Output: I will use HttpURLConnection object to establish socket connection between
the user smartphone and the server smartphone that is mounted on the quad copter.
Once the connection is established the input stream of the socket is opened. Then a
Bitmap object will be used to decode the jpeg image that is stored in the input stream
using the function BitampFactory.decodeStream(InputStream inputStream).

Task Name: Android background service for live video streaming

Accomplished By: Sifat Sultan
Starting Date: 30 March 2014
Completion Date: 15 April 2014
Obstacles Faced:

The activity in the foreground should be able to bind to the background service and
8 access all data.

Almost no documentation existed for streaming video from background.

Percentage Completion: 100%

Output: The andorid device which is mounted on top of the quadcopter can stream live
video using the 4G network, 3G network or WiFI internet.

33
 Task Name: Android background service for getting direction from compass
Accomplished By: Muhammad Obaidullah
Starting Date: 5 April 2014
Completion Date: 15 April 2014
Obstacles Faced:

Tried to implement the tilt compensated compass using the libraries provided by
researchers in the field but found out that the tilt compensation reduces the accuracy
of the smart-phone compass.
9
A smartphone compass does not provide the tilt compensated direction. It is the
raw compass direction where the value can be extremely wrong if the orientation of
the android phone itself is not correct.

Percentage Completion: 100%

Output: The andorid device which is mounted on top of the quadcopter can know the
direction of the quadcopter and is ready for performing flight calculations.

Task Name: Ground Station Windows C# application interface design

Accomplished By: Muhammad Obaidullah
Starting Date: 2 February 2014
Completion Date: 20 February 2014
Obstacles Faced:
10
Completely new to C# programming and no part of the C# application is
copied from internet. All lines of code written by Obaidullah.

Percentage Completion: 100%

Output: Base station was designed to incorporate XBee communication control.

Task Name: Embedding GPS functionality in the C# application.

Accomplished By: Muhammad Obaidullah
Starting Date: 28 April 2014
Completion Date: 23 May 2014
Obstacles Faced:

GMap library was used to embed the map into the C# application.
11 GMap provides many pre-built functions like drawing a polygon and defining GPS
point using Latitude and longitude.

Percentage Completion: 70%

Output: The ground station can be used to set way-points for the quadcopter to move
and also get the updated longitude and latitude position of the quadcopter back from the
quadcopter.

34
 Task Name: Embedding the XBee serial port in the C# application.
Accomplished By: Muhammad Obaidullah
Starting Date: 19 February 2014
Completion Date: 20 February 2014
Obstacles Faced:

12 Programming the Serial port and changing the settings for the port by using the
combo box.

Percentage Completion: 100%

Output: The user can select the port to start communicating with the XBee explorer
and select the baud rate by which the user can communicate. The data can be sent auto-
matically using the hand gestures or it can be manually entered in the text-boxes and sent.

Task Name: GPS follower background service for way-point determined flight control.
Accomplished By: Muhammad Obaidullah & Sifat Sultan
Starting Date: 25 May 2014
Completion Date: 7 June 2014
Obstacles Faced:

No previous knowledge of php and databases.

13 Figuring out the perfect closed-loop control for reaching the way-point.

Percentage Completion: 0%
Output: The user can select the waypoints on the C# application and those waypoints
will be uploaded to the database. From the database, the waypoints are grabbed by
this background service and accordingly the OBFlightPackets are generated to make the
quadcopter go to the desired way-point.

Task Name: Embedding live update of the altitude in the C# application.

Accomplished By: Muhammad Obaidullah
Starting Date: 24 February 2014
Completion Date: 25 February 2014
Obstacles Faced:

14 Getting the serial string from the XBee and performing string operations to extract
the height data.

Percentage Completion: 100%

Output: The sonar sensor which is mounted on the quadcopter can give live height
feedback to the micro-controller and XBee. The C# application shows the current height
in the nice progress bar.

35
 Task Name: Writing the Arduino C code for the AtMega328.
Accomplished By: Muhammad Obaidullah
Starting Date: 11 February 2014
Completion Date: 27 May 2014
Obstacles Faced:

Sonar sensor requires very slow triggers of about 20 Hertz. The micro-processor
cannot wait for the sonar sensor to send the sound signal, receive the signal back
and calculate the distance.
15 Drawing a clear line between the inputs and the outputs of the quadcopter system
and figuring out the finite state machine diagram.

Implementing the OBFLightPacket protocol in the C language by casting the in-

coming data into byte array and send to the ESCs.

Percentage Completion: 90%

Output: The AtMega 328 chip can be mounted on the controller board and the
quadcopter is fully functional and ready with features such as AtMega 328 to android
communication, height stabilization, XBee control, and mode select.

Task Name: Designing & updating the circuit diagram for the controller board.
Accomplished By: Muhammad Obaidullah
Starting Date: 4 February 2014
Completion Date: 26 March 2014
Obstacles Faced:

The tilt compensated compass is connected via I 2 C Master-slave configuration and

the JTAG SPI programmer header is also connected to AtMega328.

Providing the 5V logic power to the circuit board by using the ESC 3 pin headers.

Kill switch is an essential and necessary requirement especially when in initial testing
16 stages. Implemented the kill switch using the infrared receiver.

Percentage Completion: 100%

Output: The quadcopter can communicate with the tilt compensated compass and
be programmed using the JTAG SPI header directly using the .hex format binary file.
Additionally, the power is given to the circuit board using the ESCs logic output. This
improves the efficiency of the system because the ESCs have a buck converter inside to
step the voltage down from 11.1V to 5V using the PWM switching regulation technique.
Any TV remote with infrared LEDs can be used to shut the quadcopter OFF completely
as any change in the infrared signal causes and interrupt to be raised whereby the
quadcopter switches immediately to the emergency mode/state.

36
 Task Name: Soldering & mounting the equipment on the PCB
Accomplished By: Muhammad Obaidullah
Starting Date: 24 December 2013
Completion Date: 28 May 2014
Obstacles Faced:

There was a vias under the AtMega so it had to be soldered before all bracket was
to be soldered.
17
The PCB machine failed several times before a good PCB was printed. A lot of
time and effort was wasted in this. This also caused the project to be delayed.

The first versions had some faults in them so had to be printed again and again.

Percentage Completion: 100%

Output: The quadcopter PCB is completely soldered with all the features and ready to
be mounted on top of the quadcopter for autonomous flight.

10 Conclusion
The project comes with a variety of features that needs to be accomplished in the mentioned
schedule. This is made possible due to the Gantt chart that shows a detailed breakdown of how
the tasks will be approached. In our chart we tried to give a very reasonable duration to each of
the task that needs to be carried. The duration was based on various factors and experiences. For
instance, one may notice that; Printing Body; one of our sections in Gantt chart took a duration
that is about double and triple that of other parts, it was decided after a serious discussion and a
unanimous agreement from all the member that it; in reality; required a considerable amount of
time as it often demanded that the task be redone a couple of times until a desired result is reached.
Similarly there are some tasks that were given a very brief amount of time as the team believed
that we cannot afford spending a significant portion of the given time in a task that doesnt have
that amount of an impact in the later and the main portion which deserved our focus time. The
task of choosing the part resulted to be a challenging endeavor since we had to balance the quality
with the cost of the product, it must be also mentioned that a great deal of our part was ordered
from international countries. This was among other a primary obstacle that this project will face.

The team members would like to ask everyone for honest prayer so that the project can be
successful.

References
[1] Jennifer Bell. (2013) Road accidents account for almost 70% of head injuries
at one UAE hospital [Online]. Available: http://www.thenational.ae/uae/health/
road-accidents-account-for-almost-70-of-head-injuries-at-one-uae-hospital

[2] ADAFRUIT (2010) FTDI Friend Breakout Board+ (tutorial) [Online]. Available: http:
//www.adafruit.com/blog/2010/09/16/ftdi-friend-breakout-board-tutorial/

37
[3] Prospector. Polylactic Acid (PLA) Typical Properties [Online]. Available: http://
plastics.ides.com/generics/34/c/t/polylactic-acid-pla-properties-processing

[4] Mark Johnson Cutler. (2010) Design and Control of an Autonomous Variable-Pitch Quadrotor
Helicopter [Online]. Available: http://acl.mit.edu/papers/Cutler_Masters12.pdf

[5] Leap Motion Documentation. (2013) Understanding the Java Sample Application [On-
line]. Available: https://developer.leapmotion.com/documentation/Languages/Java/
Guides/Sample_Java_Tutorial.html

[6] Mando. How a Gyro Works [Online]. Available: https://learn.sparkfun.com/tutorials/

gyroscope/how-a-gyro-works

38