Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Coverage Planning for Synthetic Aperture Radar Imaging
In this project the student will implement and evaluate a coverage planner for ground penetrating synthetic aperture radar. The planner will be evaluated against existing planners and integrated and validated in real platform experiments.
Worldwide more than 60 regions and 1600km2 are contaminated with landmines and improvised explosive devices (IEDs). Contaminated territory is growing, but land release is stagnating at approximately 200km2 per year since 2011 [1]. Airborne ground penetrating synthetic aperture radar (GPSAR) could be a promising new technology to quickly and savely survey large suspected hazardous areas [2].
In order to generate high resolution radar images, different operation modes exist. A simple and effective mode is the stripmap mode where the side-looking radar is attached to the copter and moves along a linear trajectory at constant velocity. The target illumination length determines the azimuth resolution of the image. Thus long strips that illuminate the target as long as possible are desirable. An additional constraint that occurs for GPSAR is relatively low altitude between 2-5m in order to steer the signal into the ground. Last but not least the side-looking geometry enforces the copter to fly offsetted to the target.
In this project the student will extend robot coverage planning in polygon environments with no-flight-zones to account for the constraints given by GPSAR. An effective way to generate straight line coverage plans in general polygons with holes is the boustrophedon decomposition [3]. Here the polygon is decomposed into simple cells which can be covered by simple back and forth motion. Connecting all cells results in coverage of the original polygon. [4] implements and extends this approach to solve the combinatorial connection problem as a generalized traveling salesman problem to reduce transition times between cells. [5] and [6] reduce transitions and redundant sweeps even further by optimizing over each single strip. However, none of these approaches consider the side-looking geometry, optimize the target illumination time or create multiple looks for the radar.
Worldwide more than 60 regions and 1600km2 are contaminated with landmines and improvised explosive devices (IEDs). Contaminated territory is growing, but land release is stagnating at approximately 200km2 per year since 2011 [1]. Airborne ground penetrating synthetic aperture radar (GPSAR) could be a promising new technology to quickly and savely survey large suspected hazardous areas [2].
In order to generate high resolution radar images, different operation modes exist. A simple and effective mode is the stripmap mode where the side-looking radar is attached to the copter and moves along a linear trajectory at constant velocity. The target illumination length determines the azimuth resolution of the image. Thus long strips that illuminate the target as long as possible are desirable. An additional constraint that occurs for GPSAR is relatively low altitude between 2-5m in order to steer the signal into the ground. Last but not least the side-looking geometry enforces the copter to fly offsetted to the target.
In this project the student will extend robot coverage planning in polygon environments with no-flight-zones to account for the constraints given by GPSAR. An effective way to generate straight line coverage plans in general polygons with holes is the boustrophedon decomposition [3]. Here the polygon is decomposed into simple cells which can be covered by simple back and forth motion. Connecting all cells results in coverage of the original polygon. [4] implements and extends this approach to solve the combinatorial connection problem as a generalized traveling salesman problem to reduce transition times between cells. [5] and [6] reduce transitions and redundant sweeps even further by optimizing over each single strip. However, none of these approaches consider the side-looking geometry, optimize the target illumination time or create multiple looks for the radar.
- Literature review on trajectory and coverage planning for GPSAR
- Implementation of a new coverage planner for GPSAR
- Evaluation against existing planners
- Integration and validation on the real platform
- Literature review on trajectory and coverage planning for GPSAR - Implementation of a new coverage planner for GPSAR - Evaluation against existing planners - Integration and validation on the real platform
- Computational geometry
- Combinatorial optimization
- Basic C++ recommended
- [1] "Landmine Monitor 2016.", International Campaign to Ban Landmines.
- [2] Schartel, Markus, et al. "UAV-based ground penetrating synthetic aperture radar." 2018 IEEE MTT-S International Conference on Microwaves for Intelligent Mobility (ICMIM). IEEE, 2018.
- [3] Choset, Howie, and Philippe Pignon. "Coverage path planning: The boustrophedon cellular decomposition." Field and service robotics. Springer, London, 1998.
- [4] https://github.com/ethz-asl/polygon_coverage_planning
- [5] Bochkarev, Stanislav, and Stephen L. Smith. "On minimizing turns in robot coverage path planning." 2016 IEEE International Conference on Automation Science and Engineering (CASE). IEEE, 2016.
- [6] Lewis, Jeremy S., et al. "Semi-boustrophedon coverage with a dubins vehicle." 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2017.
- Computational geometry - Combinatorial optimization - Basic C++ recommended
- [1] "Landmine Monitor 2016.", International Campaign to Ban Landmines. - [2] Schartel, Markus, et al. "UAV-based ground penetrating synthetic aperture radar." 2018 IEEE MTT-S International Conference on Microwaves for Intelligent Mobility (ICMIM). IEEE, 2018. - [3] Choset, Howie, and Philippe Pignon. "Coverage path planning: The boustrophedon cellular decomposition." Field and service robotics. Springer, London, 1998. - [4] https://github.com/ethz-asl/polygon_coverage_planning - [5] Bochkarev, Stanislav, and Stephen L. Smith. "On minimizing turns in robot coverage path planning." 2016 IEEE International Conference on Automation Science and Engineering (CASE). IEEE, 2016. - [6] Lewis, Jeremy S., et al. "Semi-boustrophedon coverage with a dubins vehicle." 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2017.