Professional Documents
Culture Documents
A. System Overview
Our Aerial Surveillance System consists of a flying
machine or a drone as the primary component. This flying
machine is capable of autonomous flight and supports
custom mission planning through GPS waypoint
navigation. The mission is built on our laptop using
previously available ground stations like Mission Planner
and APM Planner. After building the mission it is
uploaded on the drone. To do this a 915MHz transmitter is
used for establishing connection between the flying
Fig.3. Radio Controller [9] machine and our laptop. Once, our flying machine is in
flight we can see all the flight data like altitude and the
B.5 915 MHz Telemetery Radio Transmitter drone’s current location on our laptop. The laptop is acting
as a ground control station to monitor the mission.
A Telemetry radio (3DR Radio Telemetry v2) of
frequency 915 MHz [9] is used for establishing A USB camera interfaced with Raspberry Pi is also
communication between quadcopter and ground station while attached to the flying machine. A wireless adapter is also
planning an automatic mission. The range of this radio is up to attached to it. The Raspberry Pi is connected to a Wi-Fi
1km (0.6 miles) [9]. Hotspot which is on the same network as our laptop. We
are connecting to the Raspberry Pi by establishing an SSH
B.6 Raspberry Pi B+ Microcontroller with wireless adapter connection so that we can give commands to it and also
Raspberry Pi is a cheap, single board, credit card- sized access its GUI from our laptop. The camera on the flying
computer. SD cards are used to store its operating system like machine which is also attached to the Raspberry Pi is live
Debian and Arch Linux ARM [11]. It has a Broadcom system streaming the video data to our laptop through this path.
on chip which consists of ARM compatible CPU, an on chip This video data is further processed.
GPU, 1 to 4 USB slots, HDMI and composite video output We can also make the Raspberry Pi play stored
and a 3.5 mm phono jack for audio. The speed of its CPU warning messages and alarms using our laptop. Thus we
ranges from 700MHz to 1.2GHz [12]. It also has an on board have a complete Aerial Surveillance System. (Fig.8.)
memory ranging from 256MB to 1GB RAM [13]. GPIO pins
are also available to provide lower level output which support
common protocols like I2C [13]. It promotes python as the
main programming language.
LOITER 1st FAIL Low Inbuilt Battery Fig.6. First flight in Standard Mode (in front of AMURoboClub,
MODE battery mechanism to charged to Mechanical Engg. Dept., Aligarh Muslim University)
signal troubleshoot 100%
Battery charge=
25%, IRIS B.2 Auto Mode
returns to its In this mode, the quadcopter was made to fly through a
launch point and series of customized waypoints. These waypoints were
lands.
designed at a ground station and then sent to the drone via a
2nd FAIL Low GPS lock Tried in open connection between the two. We experimented with various
GPS required ground with ground stations as discussed below.
signal clear sky
3rd FAIL Calibra Need Calibration Hold down
We started with Droid Planner 3, an android based ground
tion F/S button station that allows us to plan and build missions. We a built a
error simple mission successfully. Then the next step was to
connect the application with the quadcopter to upload this
RC Motors __ Calibration __
Calibra armed completed
mission on it. There were five kinds of connection possible:
tion
• USB connection through telemetry radio
4th Succes __ Flight Successful __
s Fig1 • Direct USB connection
RTL & Succes __ Successful __ • TCP connection
Landin s landing at launch
g point • UDP connection
• Bluetooth connection
We tried all of them. After a few failed attempts due to
incompatible devices that did not support features like Mobile
High Definition Link (MHL) and USB on-The- Go, and
various other reasons we finally connected to the drone using
a 915MHz transmitter. The mission was then successfully to it and powered it on. The Raspberry Pi was powered
uploaded on the drone and we had our first autonomous flight. through the drone’s battery. After that we established an SSH
Then many other missions were built and executed. Some of connection between the Raspberry Pi and our laptop using
them are shown (Fig. 11 and 12). We also executed some Putty so that we could access its GUI and command line from
problems like abrupt behavior of the quadcopter or landing a our computer [17].
little ahead than the actual land waypoint. But, overall it
worked well. Next we built a Raspberry Pi webcam server. We first
downloaded and installed Motion. Motion is a software
package that examines video signal from the camera and
detects if any change has occurred in the picture or a motion
has taken place [18]. Then we edited its configuration file
[19]. We made the following modifications:
• We switched on Daemon, a computer program that
runs in the background rather than being controlled by
the user directly. This program performs stated
operations at particular intervals or in reaction to
certain events. This is done so that Raspberry Pi
camera continues capturing video data in the
background and streaming it to an LCD monitor.
Fig.7.Autonomous Missions. • We switched off webcam localhost (Raspberry Pi) to
In order to build missions on our laptops wse then moved access the Raspberry Pi webcam from another
on to Windows based ground Station called the Mission computer.
Planner [15] and also experimented with APM Planner [16] • Next we had set the value of webcam_maxrate to 100.
which works on Linux, Windows and MAC OS. They work This enables real time streaming but requires more
for different kinds of UAVs like plane [17], copter [18] and bandwidth.
rover [19] and also support simulation software like SITL
[16]. Fig.4 shows a mission built on Mission Planner. • After that we had set the framerate to 100 so that the
webcam captures 100 frames per second for an
unruffled video stream.
• At last we had set the desired width and height of the
display.
We then switched on the daemon in the motion file for the
same reasons as stated above. Finally, we gave the command
to start the motion service and checked the video stream on
our remote desktop by accessing port 8081 at the IP address of
Raspberry Pi. This completed our second and third sub-phases
Fig.8. Mission planned in Mission Planner as given in the ‘Plan’ at the beginning of this paper. A live
streaming example is shown below. We can see the video at
After thoroughly testing the drone we concluded that it the IP address of the Raspberry Pi.
would serve as a good foundation for our aerial surveillance
system. It had the desired capabilities of handling real life
situations such as insufficient battery, low GPS signal, loss of
communication with the ground control station etc. The
automatic mission planning using waypoints was also quite
convenient and precise. It supported features like return to
launch or landing at a predefined location. So, we could now
confidently state that we had found a solid base for our project
on which new features will be added to make our desired
flying machine.
On completion of the testing phase, we began the
integration phase which is discussed in the subsequent
sections.
Fig.9. A live- streaming example
C. Interfacing Camera with Raspberry Pi and Live
Streaming Video Data to an LCD Monitor D. Enabling Alarm System
We began by setting up the Raspberry Pi. We loaded The next sub-phase was to enable the Aerial Surveillance
Raspbian, its operating system on it using 8GB class 4 SD System to play pre-recorded warning messages on detection of
card. Then we attached a USB webcam and a wireless adapter any illegal activity. The Raspberry Pi consists of phono jack
for audio. We can easily play audio messages using it. For features but our Aerial Surveillance System is relatively
this, we first checked whether our Raspberry Pi has cheap. There is also a provision to incorporate these features
snd_bcm2835 module (hardware abstraction layer) installed in the IRIS+ using gimbal & FPV kit but they are still costlier
for the audio I/O [20]. By default, the O/P is set to HDMI than our Aerial Surveillance System design. In table 2, we
audio interface if available, otherwise analog. We forced by present a cost analysis and comparison of our Aerial
the value of “n” in “amixer cset numid=3 n” command to 1. Surveillance System with already available surveillance
Here, n can be 0 (auto), 1 (analog), 2 (HDMI). Then, we drones and those which were installed using gimbal, FPV &
installed mpg321 (or omxplayer) package to play mp3 file and IRIS+ altogether.
executed it through command line [20]. The audio messaged TABLE II.
already stored in the Raspberry Pi & are then played when
desired. This completes final sub-phase of our project. DRONE PRICE
E. Assembling the Aerial Surveillance System
SOLO w/ Pre-Installed
$1149.95
Finally, we attached together all the components of our Gimbal in 3DR Backpack
Aerial Surveillance System, namely the drone, Raspberry Pi [21] IRIS+ $599.99
and camera. The Raspberry Pi was firmly fixed on the
[22] Tarot
backside of the drone and the camera was tightly fastened so IRIS+ T-2D $210.00
that it remains stationary. The Raspberry Pi was powered assembled Gimbal
$1174.98
through the drone’s battery. For this, we created our own
UBEC (universal battery elimination circuit) using LM-7085 [23]GoPro
$364.99
Hero 3
(Fig.13).This circuit converted the 12.6V that we were getting
from the drones battery to Raspberry Pi’s operating voltage [21] IRIS+ $599.99
i.e., 5.0 V The final view of our aerial surveillance system is Our
[24]
shown in Fig 15. proposed
Raspberry $35.00 $657.00
Aerial
Pi 2 (approx.)
Surveillan
ce System [25] USB
$21.80
webcam
V. CONCLUSION
Our project can be used for surveillance in civil and
military areas. It can be used for monitoring of a campus,
Fig. 10. UBEC circuit created using LM-7085 office and industrial areas by various institutions, for
monitoring the borders and in peace keeping activities by the
government and can also be used to monitor private properties
by individuals. Using them for surveillance will ensure the
safety and security of not only the citizens but the soldiers as
well. The aerial surveillance system has many advantages over
the existing methods of surveillance in our country. Some of
them are listed below:
• First of all they will be faster as the drones can
continue to fly at a high speed unhindered while a
human being, even in a high speed aero plane needs
rest.
• Secondly it will be cheaper as there is no need to
deploy and pay for human resource for this task.
Further our Aerial Surveillance System is cheaper than
Fig. 11. Final View of our Aerial Surveillance System the existing drones too.
• Thirdly this method will be much more efficient and
IV. COMPARISON WITH EXISTING WORK less erroneous because human beings are more prone to
For creating our Aerial Surveillance System, we have used committing errors than machines and in this system
IRIS+ drone kit. IRIS+ is a very simple drone that does not most of the human errors will be avoided
come with a camera. We have incorporated a camera and also
enabled live streaming of video data to a control room. We • It is much more convenient because many of the
have also enabled our Aerial Surveillance System to play pre- natural obstacles that are faced while sending soldiers
recorded messages. All these things have been assembled by to hostile terrains in order to check for potential threats
us. There are already available drones that have all these
will be easily overcome. Thus the life of both, the [3] Henri Eisenbeiss, “A Mini Unmanned Aerial Vehicle (Uav): System
soldier and the citizen is protected. Overview And Image Acquisition”, International Workshop on
"Processing And Visualization Using High-Resolution Imagery",18-20
Another use of an aerial surveillance system is to monitor November 2004, Pitsanulok, Thailand
disaster-stricken areas before taking necessary rescue and help [4] Peter M. Asaro, “The labor of surveillance and bureaucratized:killing:
measures. new subjectivities of military drone operators”, Social Semiotics, 2013
Vol. 23, No. 2, 196224,
http://dx.doi.org/10.1080/10350330.2013.777591
VI. FUTURE WORK [5] Anuj Puri, “A Survey of Unmanned Aerial Vehicles (UAV) for Traffic
Surveillance”
Our project can be further expanded. We can incorporate
[6] Rachel L. Finn, David Wright, “Unmanned aircraft systems:
collision avoidance in our Aerial Surveillance System. This Surveillance, ethics and privacy in civil applications” , comp u t e r law
can be done by using Infrared sensors to detect obstacles. & s e c u rity rev iew 2 8 ( 2 0 1 2 ) 1 8 4 e1 9 4
Thus our Aerial Surveillance System will be able to make its [7] Richard M. Thompson II,” Drones in Domestic Surveillance Operations:
path through the hindrances and reach the destination quickly. Fourth Amendment Implications and Legislative Responses,”
Congressional Research Service 7-5700 www.crs.gov R42701, April 3,
This project can also be used for other services rather than 2013
surveillance. For example, it can be used to carry light weight [8] http://www.inc.com/murray-newlands/6-drone-companies-to-look-out-
objects from one place to another thus ensuring faster for.html
delivery. We can send first aid to victims with the help of [9] https://3drobotics.com/wp-content/uploads/2015/02/IRIS-Plus-
drones. As a result, the first aid will arrive flying to the victim Operation- Manual-vH-web.pdf
in no time, while the ambulance is on its way thus saving [10] https://pixhawk.org/_media/modules/px4fmu-manual-v1.6.pdf
every minute that is crucial for the victim’s life. The doctor [11] https://www.raspberrypi.org/help/what-is-a-raspberry-pi/
will be able to see the victim through our video and also give [12] https://en.wikipedia.org/wiki/Raspberry_Pi
instructions through the alarm or audio system that plays [13] https://www.adafruit.com/datasheets/pi-specs.pdf
messages. We believe our project has a lot of potential in [14] http://ardupilot.org/copter/docs/common-choosing-a-ground-
benefitting humanity. station.html#comparison
[15] http://ardupilot.org/ardupilot/index.htm#l
ACKNOWLEDGMENT [16] http://python.dronekit.io/develop/sitl_setup.html
[17] http://www.circuitbasics.com/access-raspberry-pi-desktop-remote-
We are indebted to Mr. Sarfaraz Alam Khan for gifting the connection/
3DR IRIS+ drone-kit to AMU Roboclub, which we have used [18] http://www.lavrsen.dk/foswiki/bin/view/Motion/WebHome
in our project. We thank him profusely and appreciate his [19] http://pimylifeup.com/raspberry-pi-webcam-server/
generosity and his passion for promoting the innovative spirit [20] http://www.raspberrypi-spy.co.uk/2013/06/raspberry-pi-command-line-
amongst the students of AMU. audio/
[21] https://store.3dr.com/products/iris
REFERENCES [22] https://store.3dr.com/products/tarot-t-2d-brushless-gimbal-kit
[1] Malatone Bouasym, “USCG Search And Rescue Drone”. Engineering [23] http://www.amazon.com/GoPro-CHDHX-301-HERO3-Black-
Technology Capstone Project. Spring 2015 Edition/dp/B009TCD8V8
[2] Kenzo NANOMI, “Prospect and Recent Research & Development for [24] http://www.amazon.com/Raspberry-Pi-Model-Project-
Civil Use Autonomous Unmanned Aircraft as UAV & MAV”. Journal Board/dp/B00T2U7R7I
of System Design and Dynamics, Vol.1, No. 2, 2007 [25] http://www.amazon.com/Digital-Additions-High-Speed-
Webcam/dp/B005BPM214