Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Autonomous Radiation Mapping using mutliple UAVs
The goal of this project is to extend and further develop a tool for autonomous radiation mapping using multiple small Unmanned Aerial Vehicles (UAVs), equipped with visual, inertial and radiation sensors onboard, as well as GPS.
Unmanned systems can be considered as an ideal tool to collect data in hazardous areas without
putting humans in danger. An example of a hazardous environment is one following nuclear events,
such as after accidents or terrorist attacks with “dirty bombs”. Being able to quickly increase the
situational awareness, e.g. with respect to the actual level of contamination or the presence of human
casualties, is an important requirement to limit hazards to first responders, such as search and rescue
or decontamination teams and to increase their planning efficiency. As time and available man-power
are of the essence, employing a team of autonomous UAVs to map the radiation in an area of interest
is a promising avenue to help improving the situational awareness for first responders and subsequently
help to make informed decisions minimizing the risk to humans.
Starting of with a predefined area of interest, the UAVs should explore their environment autonomously in an
efficient manner, while coordinating their motions to avoid collisions (both with obstacles and other UAVs in the area) and
maximize their covered area per unit of time. While the UAVs estimate their local environment to allow for
an autonomous flight, the collected data can be shared with a base-station that can perform the global
mission planning and coordination of the UAVs. Once a basic (radiation) map of the area of interest has been created, this
map can be refined i.e. by high level commands from a human operator (such as gathering more samples in an uncertain region).
Finally, the visual representation of the collected data in an intuitive way is key to allow the team of rescuer to make
informed decisions based on the collected data.
This work is part of a project for civilian
protection with Armasuisse and a successful method will directly be used within in the framework
developed for this project.
Unmanned systems can be considered as an ideal tool to collect data in hazardous areas without putting humans in danger. An example of a hazardous environment is one following nuclear events, such as after accidents or terrorist attacks with “dirty bombs”. Being able to quickly increase the situational awareness, e.g. with respect to the actual level of contamination or the presence of human casualties, is an important requirement to limit hazards to first responders, such as search and rescue or decontamination teams and to increase their planning efficiency. As time and available man-power are of the essence, employing a team of autonomous UAVs to map the radiation in an area of interest is a promising avenue to help improving the situational awareness for first responders and subsequently help to make informed decisions minimizing the risk to humans.
Starting of with a predefined area of interest, the UAVs should explore their environment autonomously in an efficient manner, while coordinating their motions to avoid collisions (both with obstacles and other UAVs in the area) and maximize their covered area per unit of time. While the UAVs estimate their local environment to allow for an autonomous flight, the collected data can be shared with a base-station that can perform the global mission planning and coordination of the UAVs. Once a basic (radiation) map of the area of interest has been created, this map can be refined i.e. by high level commands from a human operator (such as gathering more samples in an uncertain region). Finally, the visual representation of the collected data in an intuitive way is key to allow the team of rescuer to make informed decisions based on the collected data.
This work is part of a project for civilian protection with Armasuisse and a successful method will directly be used within in the framework developed for this project.
1. Familiarization with the existing framework for single UAV radiation mapping as well as research into methods for multi-agent path planning and mapping in the literature.
2. Extend the existing framework to allow multi-UAV planning in a server-client architecture and test the approach in an simulated environment
3. Development of a suitable visualization for the end-user.
4. Implementation of the simulated ideas on a real platform in a realistic environment.
5. Evaluate the approach in terms of robustness and time-efficiency in a real setup.
1. Familiarization with the existing framework for single UAV radiation mapping as well as research into methods for multi-agent path planning and mapping in the literature. 2. Extend the existing framework to allow multi-UAV planning in a server-client architecture and test the approach in an simulated environment 3. Development of a suitable visualization for the end-user. 4. Implementation of the simulated ideas on a real platform in a realistic environment. 5. Evaluate the approach in terms of robustness and time-efficiency in a real setup.
- C++ programming experience
- Background in visual SLAM/sensor fusion and path planning desired
- Experience in mobile robotics, Linux, ROS are beneficial
- C++ programming experience - Background in visual SLAM/sensor fusion and path planning desired - Experience in mobile robotics, Linux, ROS are beneficial
Interested Students please write to
Yves Kompis (ykompis@ethz.ch) and Luca Bartolomei (lbartolomei@ethz.ch)
Interested Students please write to Yves Kompis (ykompis@ethz.ch) and Luca Bartolomei (lbartolomei@ethz.ch)