Autonomous Detection of Hazardous Gas Emissions using UAVs

Autonomous Detection of Hazardous Gas Emissions using UAVs
Autonomous Detection of Hazardous Gas Emissions using UAVs – Johnathon Hare
Project Description
In my proposed project, I plan to address the problem of autonomously detecting airborne gas particles using gas sensors
that are mobilized using unmanned aerial vehicles (UAVs). The results of this project will advance research in using
UAVs autonomously in different missions that are very hazardous for humans including detecting smoke and carbon
particles resulting from domestic or forest fires, detecting chemical fumes emitted by buried unexploded ordnance (UXO)
such as landmines, and detecting hazardous emissions from chemical spills. The main hypothesis I plan to investigate is
whether a commercially available, off-the-shelf gas sensor can be suitably integrated on a UAV platform to detect
ambient gas particles. The main challenges in this problem include addressing the weight constraints of the UAV’s
payload and registering a consistent reading on the gas sensor in the presence of the turbulence in the air caused by the
UAV’s rotors. To verify my hypothesis, I will undertake two major activities in this project. First, I will design a boom or
support mechanism to attach the gas sensor on the UAV using 3D printed parts so that the sensor can be in the vicinity of
the gas source, without getting the UAV too close to it. Simultaneously, the design of the boom must ensure that it does
not interfere with the landing mechanism of the UAV. After successful integration, I plan to implement an autonomous
controller program on the UAV that will use the output from the gas sensor and make the UAV hover above a gas source
and land once the gas sensor senses gas particles in the air. I will test my proposed technique with the carbon monoxide
(CO) gas sensor from RAE Systems that will be integrated on the Pelican UAV. The rest of this proposal describes the
background of my proposed research, the methodology I will use for conducting the research, and the project timelines,
roles and budget.
Background
Recently, the CMANTIC Lab (http://cmantic.unomaha.edu) at UNO has developed the COMRADE (COoperative MultiRobot Automated Detection) system for autonomous landmine detection by a team of robots [1]. The focus of the
COMRADES project was to develop autonomous coordination techniques between robots using techniques such as
swarming, flocking, task allocation and game theory (e.g., auctions). The robots in COMRADES are ground robots and
the main detection sensors are metal detectors for detecting landmines. A potential limitation of COMRADES, especially
while detecting landmines, is that ground robots can accidentally trigger a landmine explosion due to the pressure exerted
by the robot’s weight in the vicinity of the landmine [2,5]. In contrast, UAVs can enable non-contact, standoff detection of
landmines and other UXO and provide a suitable way to mitigate their accidental explosion. In [1], Dasgupta et al.
mention that one of the key findings of the COMRADES project is that the detection and location accuracy of landmines
can significantly improve, and false positives can be reduced, if multiple robots with different types of sensors can locate
landmines [3]. Using a different type of sensor on UAVs, which complements the metal detector on the ground, robots
can improve the quality of landmine detection. The first step in this direction of research is to see if detection systems can
be implemented on a UAV without impeding its ability to fly. Many UXO, including landmines, have been reported to
emit nominal, yet detectable quantities of gaseous fumes [4]. Detecting these gaseous emissions using the appropriate
sensor can provide an additional modality of detection in addition to electromagnetic detection used by the metal detectors
on ground robots. For this project, I will be working with a commercially available gas sensor called ToxiRAE Pro as the
detection system and the AscTec Pelican as the UAV. The CMANTIC lab plans on taking key concepts that were
developed in COMRADES such as multi-robot coverage and multi-robot task allocation, and develop a similar system
with autonomous UAV platforms. This project will be a step in that direction to validate if UAVs can work in large scale
multi-robot systems, to assist ground robots in completing tasks like detection of threats.
Overview
My major research will be on the analysis and integration of the ToxiRAE Pro gas sensor, shown in Figure 1, on the
Pelican UAV, shown in Figure 2. The Pelican UAV is already available in the CMANTIC lab. Software algorithms for
autonomously flying the UAV through a specified flight path and hovering at a fixed altitude have already been developed
by CMANTIC lab research. The CMANTIC lab also has a 3D printer that will be used to construct the gas sensor’s
support mechanism using 3D printed, plastic material. My first task will be to develop a support mechanism for attaching
the ToxiRAE Pro sensor on the Pelican UAV. Figure 3 shows a possible method for attaching the ToxiRAE Pro sensor
on the Pelican UAV. I would first try to slim down some of the existing parts on the Pelican UAV, as some hardware is
attached asymmetrically and might cause instability of the UAV if I add the proposed structure as is on the UAV. I plan to
attach the ToxiRAE Pro sensor to the lowest part of the Pelican UAV so that the clearance of the UAV’s rotors from the
gas source is increased, which would reduce the dissipation of the fumes from the rotors’ turbulence. Depending on how
ToxiRAE Pro works, I would incorporate a tube and a suction device (design to be determined), that will suspend
downwards from the UAV. Once this design is complete, I will explore techniques for hovering and landing the Pelican
UAV with the added weight of the ToxiRAE Pro sensor and its associated support mechanism. I plan on developing a
Autonomous Detection of Hazardous Gas Emissions using UAVs – Johnathon Hare
program in ROS (Robot Operating System) that would enable autonomous hovering and landing of the Pelican UAV with
the integrated ToxiRAE Pro sensor. The data for the autonomous hovering program would be obtained from an ultrasonic
sensor on the Pelican UAV that points at the ground and gives real-time altitude data of the UAV, and a gyroscope sensor
that gives the orientation of the UAV with respect to the ground. These data will be used to control the altitude and
orientation of the UAV and make it land gracefully near the gas source, after the gas sensor detects it.
Figure 1. ToxiRAE Pro personal gas detector.
Figure 2. Asctec Pelican UAV
Figure 3. Proposed concept of support
mechanism for gas sensor on UAV.
Methodology
First, I will learn how the ToxiRAE Pro sensor works and how to enable wireless communication between the ToxiRAE
Pro sensor and the Pelican UAV. Also, in this phase, I will test the ToxiRAE Pro sensor’s accuracy with carbon monoxide
(CO) emitted from a combustion engine like a lawn mower or a car (the sensor specifications mention that it has a particle
resolution of 500 ppm with a resolution of 1 ppm for CO particles, which would enable us to detect CO with high
accuracy). After this initial testing I will begin drafting and design methods for the support mechanism to attach the
ToxiRAE Pro sensor. The ToxicRAE Pro sensor weighs 220g, while a UAV’s maximum payload is 450g. The support
mechanism I will design will have to remain within the weight and dimension limitations of the UAV and also allow the
UAV to take off and land without intrusion. Initial prototypes will be made with inexpensive material like pressed
paperboard cutouts to determine how the structure will fit on the UAV. Once a suitable paperboard design is identified, I
will proceed to design a support mechanism on 3-D printed material. If it is too expensive to print all of the parts I would
buy small things to reinforce the structure of the 3-D printed parts. Examples of such reinforcing parts would be small
tubing, screws, nuts, and bolts. Once everything is assembled, I will begin developing a software program that will make
the Pelican UAV autonomously hover with the attached sensor. Once this program is complete I will implement and test a
program that would make the Pelican UAV fly to a CO source, hover at the location of the source to collect more samples,
then once it successfully detects the CO particles using the ToxiRAE Pro sensor, it will proceed to land to demonstrate
that the UAV can autonomously react to data collected by the gas sensor in real-time.
Project Timeline
Jan 1 – Feb 28, 2016
Mar 1 – Jun 30, 2016
Jul 1 – Sep 30, 2016
Oct 1 – Dec 31, 2016
Jan 1 – Jan 31, 2017
Acquire ToxiRAE Pro sensor and test its accuracy, sensitivity and range. Enable the wireless
communication between the Pelican UAV and the sensor.
Prototype methods of support mechanism for ToxiRAE Pro sensor on Pelican UAV.
Finalize design of support mechanism with 3-D parts and other material.
Assemble ToxiRAE Pro sensor on Pelican UAV.
Implement program that enables autonomous hovering and landing of the UAV using the ultrasonic and
gyroscope sensors, including the ToxiRAE Pro sensor attached with the support mechanism, but not using
data from the gas sensor.
Implement program that incorporates data collected by ToxiRAE Pro sensor from a gas source and makes
pelican hover and land near the source.
Write report and prepare presentation for RCA 2017 fair.
Student/Faculty Mentor Roles
I will investigate different approaches to integrate the ToxiRAE Pro gas sensor, and implement techniques to control the
UAV for detecting, hovering and landing processes. The faculty mentor, Dr. Jose Baca, will have weekly meetings to
guide me through the sensor integration, control of the UAV, and how to properly use all of the software. I will be in the
CMANTIC Lab in PKI while conducting the research project and for system’s validation we will execute outdoor
experiments.
Autonomous Detection of Hazardous Gas Emissions using UAVs – Johnathon Hare
Budget Justification
Dollar amount Requested
Description
$1,500
Stipend for student work- 10 hours/week, 15 weeks at $10 per hour for
a total of 150 hours
Gas Detection System:
 One ToxiRAE Pro w/ CO sensor: $620
 RAE software CD: $80
 RAE license: $300
$1,000
References
[1] P. Dasgupta, J. Baca, K. R. Guruprasad, A. Munoz-Melendez, J. Jumadinova, "The COMRADE System for MultiRobot Autonomous Landmine Detection in Post-Conflict Regions," Journal of Robotics, Hindawi, 2015.
[2] R. Albright. Cleanup of chemical and explosive munitions: location, identification and environmental remediation.
William Andrew Publishers, 2011.
[3] D. Butler, E. Caspedes, C. Cox, W. Wolfe. Multisensor methods for buried unexploded ordnance detection,
discrimination and identification. US Army Corps of Engineers, Technical Report SERDP 98-10, September 1998.
[4] J. Brynes. Unexploded Ordnance Detection and Mitigation. Springer Science & Business Media, 2008.
[5] L. Beran, B. Zelt, and S. Billings. Detecting and Classifying UXO. The journal of ERW and mine action, Spring
2013, 57-63.
Letter of Mentor Support
Office ir Research and Creative Activity
Eppley Administration Building, 203
University of Nebrasla at Omaha
Subject: Fund for Undergraduate Scholary Experiences
I am writing this letter to support Johnathon Hare’s FUSE proposal, “Autonomous Detection of
Hazardous Gas Emissions using UAVs”. I am a post-doctoral researcher and assistant director in the
CMANTIC Lab since 2012. I received my PhD degree in Robotics and Automation from Universidad
Politecnica de Madrid, Spain and have published several papers and made significant contributions to the
CMANTIC lab projects through my research on robotics. I am currently doing research in the fields of
modular robotics, unmanned autonomous vehicles, surgical robotics, and educational robotics.
This proposal attempts to advance research in using Unmanned Aerial Vehicles (UAVs) autonomously in
different missions that are very hazardous for humans including detecting gas, smoke, carbon particles
resulting from domestic or forest fires, detecting chemical fumes emitted by buried unexploded ordnance
(UXO) such as landmines, and detecting hazardous emissions from chemical spills. This work addresses
the problem of detecting airborne gas particles using gas sensors that are mobilized using UAVs. The
goals can be classified in two main parts. First, Johnathon proposes to investigate strategies to integrate a
mechanism for attaching the gas sensor on the UAV (Pelican robot, quadcopter) that satisfies the payload
limitations of the system, and stability. The mechanism should not interfere with the quadcopter during
the hovering and landing processes. Secondly, he plans to analyze and determine the correct location of
the sensor to avoid the dissipation of the fumes due to the turbulences generated by the UAV’s rotors. He
is planning to validate the overall research project by using real robots and sensors. I find Johnathon Hare
suitable for this research project since he is highly interested in robotics, particularly in the design of
mechanisms based on analysis of real problems. He also joined the UN Robotics association, where he
share and develop robotic ideas. A FUSE award will benefit his educational experiences.
Sincerely,
Jose Baca, Ph.D.
Research Associate,
CMANTIC Lab
Computer Science Department, College of IS&T
University of Nebraska at Omaha, USA
E-mail: jbacagarcia@unomaha.edu
Phone: 402.554.6043
URL: https://sites.google.com/site/bacajose/
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Related manuals

Download PDF

advertising