This is the author’s version of a work that was submitted/accepted for pub-lication in the following source:

Kersnovski, Tyler, Gonzalez, Felipe, & Morton, Kye(2017)A UAV system for autonomous target detection and gas sensing. In2017 IEEE Aerospace Conference, 4-11 March 2017, Yellowstone Confer-ence Center, Big Sky, Montana.

This file was downloaded from: https://eprints.qut.edu.au/102950/

c© 2017 IEEE

Personal use of this material is permitted. However, permission to reprint/republish thismaterial for advertising or promotional purposes or for creating new collective works forresale or redistribution to servers or lists, or to reuse any copyrighted component of thiswork in other works must be obtained from the IEEE.

Notice: Changes introduced as a result of publishing processes such ascopy-editing and formatting may not be reflected in this document. For adefinitive version of this work, please refer to the published source:

1

A UAV System for Autonomous Target Detection and Gas

Sensing

Tyler Kersnovski Queensland University of

Technology 2 George St, Brisbane City QLD

4000 tyler.kersnovski@connect.qut.edu.au

Felipe Gonzalez Australian Research Centre for

Aerospace Automation 22 Boronia Rd, Brisbane Airport QLD

4009 felipe.gonzales@qut.edu.au

Kye Morton Queensland University of

Technology 2 George St, Brisbane City QLD

4000 kye.morton@hdr.edu.au

Abstract—Monitoring of environmental gases is a demanding

task that can require long periods of observation and large

numbers of sensors. Unmanned Aerial Vehicles (UAVs)

presently represent a suitable alternative to monitor large,

remote, and difficult to access areas. Due to the wide range and

diversity of shape and size, UAVs now possess the capability of

carrying specialized sensor modules that can accurately monitor

gas concentrations. This paper describes the design and flight

testing of a UAV which carries an on-board camera and a

carbon dioxide gas sensor and is capable of autonomous gas

sensing while simultaneously visually detecting predefined

targets placed at locations inside a room. The detection system

autonomously navigates around the flight area using a waypoint

navigation system and hovers for 10 seconds once the target has

been visually identified taking Air Quality (AQ) samples as it

does so. Laboratory, bench and field test results demonstrate the

capability of the UAV to detect targets placed in a room and

analyse the air quality of samples taken during flight above the

target. The data collected during the flight is transmitted, in real

time, back to a Ground Control Station (GCS) for visualisation

and analysis, where 3D mapping of the target location and gas

concentration is presented via a web interface. Test cases verify

the capability of the subsystem’s integration and the operation

of the UAV system as a whole. The image processing algorithm

used for the target detection, based on a cascade method, proved

to be dependent on the frame transmission rates, which can

make the software considerably slower. The developed system

provides an effective monitoring system and can be used in a

wide range of applications such as gas leaks, fires, and mining

applications and if converted for use in an outdoor environment,

applications such as agriculture biomass burning emissions and

chemical and biological agent detection studies. The system

could be used in conjunction with ground sensors and integrated

into an extensive gas monitoring system.

Keywords: UAV gas sensing, Autonomous target detection,

CO2, Aircraft Navigation, Carbon Dioxide, Gas, Image Edge

Detection, Image Recognition, Object Detection, Unmanned

Aerial Vehicles, Sensor Systems.

TABLE OF CONTENTS

1. INTRODUCTION ..................................................... 1 2. SYSTEM ARCHITECTURE ...................................... 3 3. FLIGHT TESTS ....................................................... 5 4. RESULTS OF TEST CASES ...................................... 6 5. CONCLUSIONS ....................................................... 7 6. APPENDICES .......................................................... 8 7. ACKNOWLEDGEMENTS ....................................... 10 8. REFERENCES ....................................................... 10 9. BIOGRAPHY ......................................................... 12

1. INTRODUCTION

The current increases in the availability of small commercial

Unmanned Aerial Systems (UAS) has led to the development

of many different applications. These applications extend to

the development of many different types of UAVs, both fixed

wing and multirotor platforms, for the purpose of aerial

monitoring, object detection, and image processing.

"Significant research has been conducted on aerial robotic

platforms [7, 13, 15 ] as well as expanding the area of

operations for these small-scale UAVs for remote sensing [2,

3, 4, 5, 14]. One such application is Aerial Monitoring of Air

Quality (AMAQ) using target detection and image processing

to locate a target source for the air samples [6, 4, 10, 11].

Research on UAVs path planning, design and remote sensing

is an active field of research [1, 3, 4, 6, 9, 10, 11, 12].

Malaver et al [14] for example describes an AMAQ system

implemented on a fixed wing UAV platform equipped with

carbon-dioxide and methane metal oxide (MoX) and non-

dispersive infrared sensors. The system is integrated with a

network of ground sensors which send information to a

central node in real time for 3D mapping of the data obtained.

The system incorporates solar power into the design, with a

new solar panel encapsulation method used to provide power

to the UAS.

Baxter et al [3] describes another AMAQ system for a

multirotor UAS also equipped with air quality sensors. The

system is designed as a stable platform, with flight times of

up to half an hour, to provide a platform for the measurements

of ozone levels in the atmosphere. The system tests two

different sensors and is designed to possibly replace fixed or

tethered monitoring systems.

Montes et al [15] developed and tested two different plume

tracking algorithms by studying male tobacco hornworm

moth flight patterns and plume tracking methods. The paper

describes a memory based and a gradient based tracking

algorithm used to implement 2D and 3D plume tracking for a

UAV platform.

2

Figure 1: Concept of Operations

The system described in Bartholmai et al [2] follows the

development of a small UAV system integrated with gas

sensors to be used as a mobile gas sensing unit in potentially

hazardous environments such as chemical or hazardous

material spills or events such as a terrorist gas attack. The

report also describes a method of source tracking

implemented into the system.

Bretschneider et al [5] describes a small sized UAV system

utilising an optical wavelength camera to track unburied

pipelines using an on-board tracking algorithm. The UAV is

fitted with various laser and electro-optical sensors to track

the gas levels in the immediate area surrounding the pipeline

The development of a hexacopter fitted with a metal oxide

array controlled by a 32bit microcontroller running at 32

MHz for the application of gas leakage localisation is

described by Rossi et al [20]. The system develops a custom-

made sensor arrangement that allows the concurrent

evaluation of 8 different compounds.

Bing [4] presents the methods and experimental results of

three-dimensional gas distribution mapping with a micro

drone. A 3D kernel extrapolation distribution mapping

algorithm is implemented on the micro-drone’s on-board

computer. The gas distribution map is updated in real-time as

the micro-drone flies along a predefined sweeping path.

Results show that the concentration grid map obtained can

represent the real gas distribution to a certain extent, and

demonstrates the gas distribution mapping with a micro-

drone.

Neumann et al [16] proposes a solution for gas source

localisation, including source declaration. A novel pseudo-

gradient-based plume tracking algorithm and a particle filter-

based source declaration approach is presented, and applied

on a gas-sensitive micro-drone. The performance of the

proposed system in simulations and real-world experiments

is compared against two commonly used tracking algorithms

adapted for aerial exploration missions.

Peng et al [17] presents a real-time aerial gas monitoring

system for use in air quality control. The proposed method is

used to integrate a gas sensing module with a quadrotor UAV

with gas data being wirelessly transmitted to a ground station.

The system also uses image processing to automatically

identify targets for detection.

The remainder of the paper is organised as follows. Section

II describes the system architecture of the UAS with the

system broken down into subsystems and a detailed

description of the components and software developed for the

UAS is provided. Section III discusses the flight testing of the

UAV conducted in a controlled environment and the systems

performance is then evaluated. A summary of conclusions is

provided in section IV.

The live video feed from the UAV is sent to the GCS to be

processed for target acquisition. The AQ sensor data is

transmitted to the GCS and embedded with positional data. A

web interface displays the live video feed from the camera

and sensor and positional data. This enables 3D mapping of

the sensor data over the desired target.

Figure 2: AMAQ UAV

3

2. SYSTEM ARCHITECTURE

Figure 2 shows the system architecture. This overall system

can be broken up into smaller sections to provide a detailed

design.

1. Airframe, Power and Propulsion

The four principle components for this subsystem are the

airframe, motors, electronic speed controllers, and battery.

The airframe is based on a DJI F450 Flamewheel airframe.

This airframe is made from a PA6+30GF nylon 6/6 glass

fibre reinforced material providing strength and durability. A

high strength PCB board has been incorporated into the

airframe design to provide a central power distribution.

The motors used are Sunnysky X2212 motors rated at 980kV

with a maximum current draw of 17.2A under a static load.

With a voltage of 11.1 volts supplied by the battery the

motors are capable of providing 960g of thrust per motor

spinning at a maximum 5830rpm.

The Electronic Speed Controllers (ESC) used are Turnigy

Plush 25A speed controllers. These ESC’s provide enough

power and current to the motors while keeping an acceptable

buffer between the actual current draw of the motors and the

maximum current draw through the ESC’s.

The source of power for the components on the UAV is a

Multistar 3-cell 11.1V 5200mAh lithium-polymer (LiPo)

battery. The battery has a maximum capacity of 5200mAh

and produces a constant discharge rate of 5.2A and maximum

discharge rate for a period of 10 seconds of 10.4A. The

battery weighs 331g and is arranged in a 3-cell 11.1V

configuration.

Figure 3: VICON Control Diagram for UAV

2. Navigation and Autopilot

There are a number of principle components for the

navigation and autopilot system of the UAV.

The autopilot used for the flight controller of the UAV is a

3D Robotics (3DR) Pixhawk autopilot. This autopilot is an

open source autopilot reprogrammable with different

software packages to suit the application of the user. The

software used in the Pixhawk autopilot is the APM 2.6

software package uploaded onto the autopilot. This software

is used for compatibility with the Vicon and the Robotic

Operating System (ROS) software used for navigation. The

autopilot acts as a simple rate controller on the UAV.

The telemetry link is used to transfer control information

between the GCS and the UAV. The telemetry link used for

this design is a RFD900, running at 900Mhz with a baud rate

of 115200.

Figure 4: Infrared Marker

The main component of the navigation and flight control of

the UAV is a combination of Vicon motion capture system,

and ROS control software. Infrared detection markers, seen

in figure 4, are placed on the UAV airframe to allow the

Vicon system to determine the exact position of the UAV in

the flight area. This then sends the positional and attitude

information at a rate of 100Hz to the ground station computer,

as shown in figure 3.

Figure 5: Ground Station Data Flow

There are a number of processes occurring inside the GCS.

The GCS receives transmitted Vicon data and converts this

data into usable information, i.e. positional and attitude

information. This positional data is then sent to a PID

controller which first compares the position and orientation

of the UAV with the desired position and orientation then

runs a stability PID to maintain the stability of the UAV.

From this process, control commands are generated in the

form of rotational goal rates, which would move the UAV to

the desired position and orientation. The rates are sent via the

telemetry link, to the UAV autopilot which outputs the

commands to the motors. The data flow for this process can

be seen in figure 4.

4

Figure 6: Navigation Software Flow Diagram

The navigation is conducted via a python script. The pseudo

code for navigation is shown in Figure 5. The script takes in

a text file that contains waypoints for the navigation around

the search area. These coordinates are then sent to the PID

controller software which passes the control inputs onto the

autopilot. The python script is also connected via a socket

connection through a Wi-Fi modem to the image detection

software. The image detection sends messages once the target

is found which pauses the operation of the python script. The

UAV hovers for 10 seconds over the target whilst taking air

samples with the CO2 sensor.

3. Target Recognition and Acquisition

The algorithm used for target detection is detailed as follows:

The algorithm for image recognition employs two different

methods of detection for targets. The image processing

program first converts the captured images to grayscale to

reduce outside noise in the image. A circle detection program

is then applied to the images to determine the presence of

circular targets. If a circular target is found the processing

centre will highlight the circle in the image and notify the

GCS that a target has been found. If no target is found the

image is converted to a Hue, Saturation and Value (HSV)

colour range. The target colour range values are then set and

a mask is applied to the image to isolate the target colours.

Edge detection is then applied to the masked image to

determine whether the correct target shape is present. If the

target shape is present, the processing centre sends a message

to the GCS. If no target is present, there is no message sent

so the UAV continues its flight path. The software flow

diagram is presented in Figure 6.

Figure 7: Algorithm Software Flow Diagram

4. Air Quality Data Acquisition

The two components of the air quality data acquisition and

transmission systems are a CO2 sensor and a portable

microcomputer. The CO2 sensor used is a SprintIR sensor. It

is a low power and high speed response CO2 sensor

manufactured by CO2Meter. The sensing method is non-

dispersive infrared (NDIR) absorption, with warm-up time

less than 10s, accuracy of +/- 3%, using 3.25V to 5.5V of DC

power and 35mW consumption to provide a 0-5000ppm

measurement range [7]. A Raspberry Pi (RPi) 2 is used as the

microcomputer. The RPi has a CPU of 900MHz quad core

Cortex-A7, RAM of 1GB and consumes 4W of power [8]

used to process and transmit data.

The CO2 sensor output pins are connected to the RPi

microcomputer input pins through jumper wires. The RPi

also provides regulated power to the CO2 sensor while the

on-board battery is used to supply power to the

microcomputer. The RPi uses a Wi-Fi module to send the live

CO2 data back to the GCS. For the purpose of receiving and

sending the CO2 sensor data via Wi-Fi, a simple python script

is used. This is called the TCP/IP server and not only

commands reading the serial data through the RPi input pins

but also sends this data via TCP/IP sockets to the GCS

through a router. The computer located in the GCS is used to

run a python client script which accesses the data sent to the

router. This script receives the TCP/IP socket messages sent

from the RPi and writes the data into a MySQL database.

5

5. Web Interface

The web interface displays the images sent from the on-board

camera, the CO2 sensor readings and current UAV location

(x, y and z coordinates inside the flight area) in real time.

Once the images from the UAV on-board camera are sent to

the GCS router, the web page displays this data. We

developed two solutions for imagery display. One of them

accesses the streamed data from an IP camera app available

in the appropriate IP address. The other solution uses the RPi

to process the on-board camera images and store them in the

database to posteriorly access them in the web page.

Figure 8: Web Visualization Data Flow Diagram

The UAV position is retrieved from the ROS system using a

python script. The software is composed of a subscriber to

the topic created for the UAV and a TCP/IP client is run to

receive the position data and write it in the database for

further access from the webpage.

The position and air quality data are also displayed in the

webpage using PHP to interact with the database. Java Script

is used to continually updates the page as the CO2 readings

and x, y and z are also updated. MATLAB code is used to

connect to the database and retrieve the CO2 measurement,

X and Y position data in and display them in the format of a

graph. This graph is also displayed on the webpage.

A data flow of the web visualization subsystem is

summarized and presented in Figure 7. The web visualisation

can be seen in Appendix A1 and A2.

Figure 9: Search Pattern Flight Path

3. FLIGHT TESTS

The AMAQ system was subjected to two flight test cases in

a controlled flight environment.

1. Test Case 1

The first test case involves autonomously flying the UAV in

a programmed search pattern around the room, shown in

Figure 8, with the aim of continually sampling, searching,

and tacking the location of the UAV within the flying area.

The samples are transmitted to the GCS for analysis and

display on the web server.

The flight area is then broken into an X/Y quadrant grid with

the centre point of the flight area set to (0,0).

2. Test Case 2

The second test case involves placing one of each different

target at two different locations inside the flight area as shown

in Figure 1.

Targets A and B were placed inside the flight area at the

locations (0, -1) and (0, 1) respectively. The task of the UAV

was to locate these targets. Once a target has been identified

the UAV hovers above its location for 10 seconds taking air

samples and then proceeds to continue its search. This is

repeated for each target in the flight area until the end of the

navigation plan.

6

4. RESULTS OF TEST CASES

1. Test Case 1

Figure 10: Travelled path for Test Case 1

The position of the UAV inside the flight area for Test Case

1 was tracked during the flight by the Vicon system and is

presented in Figure 9 by the blue line. The planned flight path

from Figure 8 is also presented by the red line so as to see the

contrast between the planned and the actual flight paths.

Figure 11: 3D graph for travelled path and way points in

Test Case 1

A three-dimensional graph of the flight path is shown in

Figure 10; the red dots represent the mission waypoints. The

system for UAV flight also includes a precision margin of

25cm for reaching the waypoints which can be seen by the

black circles. The path travelled is shown once again by the

blue line.

The webpage continuously displayed imagery from the on-

board camera as well as the CO2 readings from the air quality

sensor and the real-time position of the UAV throughout the

flight. A screenshot of the webpage is shown in Appendix A1

and A2.

2. Test Case 2

Results for Test Case 2 show the operation of the UAV. The

results also represent the test and validation of the image

processing and the web visualisation subsystems separately

as the airframe, power and propulsion, navigation and

autopilot, and air quality components were already

successfully tested in Test Case 1.

Figure 12: Image processing code: Target A recognition

Figure 13: Image processing code: Target B recognition

For this test, the UAV was flown over both targets, A and B.

The image processing system was able to identify both of

them during the flight as shown in figures 12 and 13.

Some screenshots were taken during this flight test in order

to show the webpage operation and they are presented in

Appendix A1 and A2. These figures verify that the webpage

system was able to collect, store and display the UAV on-

board data, including the real-time images from the on-board

camera, the current carbon dioxide reading from the on-board

sensor and current UAV location as well as a graph with the

CO2 measurements and location history throughout the

flight.

7

5. CONCLUSIONS

As UAVs are now able to carry on-board high resolution

cameras and specialized gas sensors, they present a solution

to an increasing demand of both remote target detection and

air quality sensing in difficult access areas. This article

describes and validates a system which explores both remote

target detection and air quality sensing using UAVs.

The UAV tested was able to fly autonomously, sample CO2

concentrations and detect two predefined targets in an indoor

flight area. Test cases 1 and 2 verify the capability of the

subsystem’s integration and the operation of the UAS as a

whole.

The image processing algorithm described proved to be

dependent on the frame transmission rates, which can

considerably slow down the software. Beyond that, the next

step for this system is to test other image processing solutions

on-board the UAV.

For future research this project can be extended to an outdoor

solution if proper modifications are made. To do so, changes

in telemetry, image processing, and navigation need to be

considered so data analysing and storage, as well as decision

making are migrated to an on-board system.

8

6. APPENDICES

Appendix A1: Webpage screenshot – Target A

9

Appendix A2: Webpage screenshot – Target B

10

7. ACKNOWLEDGEMENTS

We would like to acknowledge Victoria Tobin, Min Shi,

Kelvin Bright and Bryce Kelly for their development and

support on the target detection software. Kane Nicholson for

development and support on the webpage design.

8. REFERENCES

[1] Al-Sabban, W. H., Gonzalez, L. F. & Smith, R. N.,

(2013). Wind-energy based path planning for

unmanned aerial vehicles using Markov decision

processes. In proceedings of 2013 IEEE International

Conference on Robotics and Automation (ICRA).

[2] Bartholmai, M., & Neumann, P. (2010). Micro-drone

for gas measurement in hazardous scenarios via remote

sensing. BAM Federal Institute of Materials Research

and Testing, Selected Topics in Power systems and

Remote Sensing.

[3] Baxter, R. A. Bush D. H. “Use of Small Unmanned

Aerial Vehicles for Air Quality and Meteorological

Measurements” in Proceedings of National Ambient Air

Monitoring Conference 2014.

[4] Bing, L., Qing-Hao, M., Jia-Ying, W., Biao, S., & Ying,

W. (2015, July). Three-dimensional gas distribution

mapping with a micro-drone. In Control Conference

(CCC), 2015 34th Chinese (pp. 6011-6015). IEEE.

[5] Bretschneider, T., & Shetti, K. UAV-based gas pipeline

leak detection. (2010) In Asian Conference on Remote

Sensing, At Nay Pyi Taw, Myanmar

[6] CO2 Meter. “SprintIR™ 20Hz High-Speed & Low-

Power Carbon Dioxide Sensor” 2015.

http://www.co2meters.com/Documentation/Datasheets/

SprintIR-Ambient-CO2-Sensor-Datasheet.pdf

[7] Doherty, P.; Granlund, G.; Kuchcinski, K.; Nordberg,

K.; Sandewall, E.; Skarman, E.; and Wiklund, J.

(2000). The WITAS unmanned aerial vehicle project.

In Proceedings of the 14th European Conference on

Artificial Intelligence, 747–755. Project URL:

http://www.liu.se/ext/witas.

[8] Fraundofer, F., (2015, July). Drone Vision – Computer

vision algorithms for drones. In IROS 2015 Conference

Precedings. IEEE

[9] Gonzalez, LF. (2005). Robust evolutionary methods

for multi-objective and multidisciplinary design

optimisation in aeronautics. In collections of the School

of Aerospace, Mechanical and Mechatronic

Engineering

[10] Lee, D. S., Gonzalez, L. F., Periaux, J., Srinivas, K. &

Onate, E. (2009). Hybrid-Game Strategies for multi-

objective design optimization in engineering. In

Journal Computers & Fluids Volume 47, Issue 1,

August 2011, Pages 189–204

[11] Lee, D. S., Gonzalez, L. F., & Whitney, E. J. (2007).

Multi-objective. From Multidisciplinary Multi-fidelity

Design tool: HAPMOEA-User Guide.

[12] Lee, D., Periaux, J. & Gonzalez, F., (2009). UAS

mission path planning system (MPPS) using hybrid-

game coupled to multi-objective optimiser. In

proceedings of ASME 2009 International Design

Engineering Technical Conferences and Computers

and Information in Engineering Conference

[13] Li, X., Yang, L., (2012, August). Design and

implementation of UAV intelligent Aerial Photography

System. Intelligent Human-Machine Systems and

Cybernetics (IHMSC), 2012 4th International

Conference on (Volume:2)

[14] Malaver Rojas, J. A., Gonzalez, L. F., Motta, N., Villa,

T. F., Etse, V. K., & Puig, E. (2015, September). Design

and flight testing of an integrated solar powered UAV

and WSN for greenhouse gas monitoring emissions in

agricultural farms. In 2015 IEEE/RSJ International

Conference on Intelligent Robots and Systems (Vol. 1,

No. 1, pp. 1-6). IEEE.

[15] Montes, G., Letheren, B., Villa, T., & Gonzalez, F.

(2014). Bio-inspired plume tracking algorithm for

UAVS. ACRA 2014 Australasian Conference on

Robotics and Automation, 2-4

[16] Neumann, P. P., Hernandez Bennetts, V., Lilienthal, A.

J., Bartholmai, M., & Schiller, J. H. (2013). Gas source

localization with a micro-drone using bio-inspired and

particle filter-based algorithms. Advanced

Robotics, 27(9), (pp. 725-738).

[17] Peng, C. C., & Hsu, C. Y. (2015, June). Integration of

an unmanned vehicle and its application to real-time gas

detection and monitoring. In Consumer Electronics-

Taiwan (ICCE-TW), 2015 IEEE International

Conference on (pp. 320-321). IEEE.

[18] Pounds, P., Mahony, R., Hynes, P., & Roberts, J. M.

(2002) Design of a four-rotor aerial robot. In Friedrich,

Werner & Lim, Patrick (Eds.) Proceedings of the 2002

Australasian Conference on Robotics and Automation

(ACRA 2002), Australian Robotics & Automation

Association, Auckland, New Zealand, pp. 145-150.

[19] Raspberry Pi Foundation. “Raspberry Pi 2 Model B”

n.d.

11

[20] Rossi, M., Brunelli, D., Adami, A., Lorenzelli, L.,

Menna, F., & Remondino, F. (2014, November). Gas-

Drone: Portable gas sensing system on UAVs for gas

leakage localization. In SENSORS, 2014 IEEE (pp.

1431-1434). IEEE.

[21] Yang, S., Zhao, K., (2015) DJI Drone Platform and

Navigation Technologies Research. In IROS 2015

Conference Precedings. IEEE

12

9. BIOGRAPHY

Tyler Kersnovski is

currently completing his

final year of study at the

Queensland University of

Technology to obtain a

Bachelor of Engineering

majoring in Aerospace

Avionics. He is a member

of RAeS and a student

member of Engineers

Australia and IEEE. He is

currently completing an internship at Hawker Pacific

Australian Avionics based in Cairns.

Felipe Gonzalez is an

Associate Professor in the

Science and Engineering

Faculty and team Leader

for Integrated Intelligent

Airborne Sensing

Laboratory at the

Queensland University of

Technology. He holds a

BEng (Mech) and a PhD

from the university of

Sydney. His research explores bioinspired optimization,

uncertainty based UAV path planning and UAVs for

environmental monitoring. He leads the CRC Plant

Biosecurity Project Evaluating Unmanned Aerial Systems

for Deployment in Plant Biosecurity and the Green Falcon

Solar Powered UAV. Felipe is a Chartered Professional

Engineer and member of professional organizations

including the RAeS, IEEE and AIAA.

Kye Morton is currently

studying a doctorate of

philosophy at the

Queensland University of

Technology (QUT). He

holds a Bachelor of

Engineering in Aerospace

and Avionics from QUT in

2014. His current

developments are towards

control methods for

complex UAV designed at

ARCAA and QUT.