Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Improved UAV detection and tracking with radar
[THE PROJECT IS TAKEN]
Small size UAVs pose a challenge in terms of their detection and tracking due to their small size. In this context, the task is to implement algorithms for enhancing the radar/vision system capabilities, allowing the detection of UAVs at larger distances and/or under poor environmental conditions.
The continuous increase of small-sized unmanned aerial vehicles (UAV) represents a threat for civilian and military activities. Moreover, small size UAVs pose a challenge in terms of their detection and tracking, especially in non-ideal weather conditions. Radar aided systems offer a robust solution for UAV target detection and tracking applications, due to their long range measurement capabilities in the presence of rain, dust and air contamination. For the case of low altitude, slow motion, small size (LSS) UAVs, the target radar signature (received power from the UAV target) is small and the radar systems receive indirect back scattered signals from the ground. Therefore, tracking such vehicles is more demanding and requires additional effort in processing the radar measurements. In this context, the task of the student is to implement and modify state of the art algorithms for enhancing the radar system component capabilities, allowing the detection of UAVs at larger distances and/or under poor environmental conditions.
In particular, this includes:
- Analysis and modelling of the expected false alarm and UAV target detection statistics.
- The implementation of a CFAR (Constant False Alarm Rate) algorithm with adaptive detection threshold matched to the expected false alarm statistics.
- The implementation of a radar based tracker, capable of initiating and maintaining track of a single UAV in adverse environmental conditions.
In a first step, the student will investigate several detection methods and will choose the best method for allowing UAV detection. The student will develop an algorithm for enhancing CFAR detection performance. Recent state of the art multi-target tracking algorithms based on Multi-Bernoulli Filters will be implemented and tested in terms of their abilities to initiate and maintain track of UAVs under varying distance and environmental conditions.
The developed algorithms will ultimately form part of a vision/radar system.
The continuous increase of small-sized unmanned aerial vehicles (UAV) represents a threat for civilian and military activities. Moreover, small size UAVs pose a challenge in terms of their detection and tracking, especially in non-ideal weather conditions. Radar aided systems offer a robust solution for UAV target detection and tracking applications, due to their long range measurement capabilities in the presence of rain, dust and air contamination. For the case of low altitude, slow motion, small size (LSS) UAVs, the target radar signature (received power from the UAV target) is small and the radar systems receive indirect back scattered signals from the ground. Therefore, tracking such vehicles is more demanding and requires additional effort in processing the radar measurements. In this context, the task of the student is to implement and modify state of the art algorithms for enhancing the radar system component capabilities, allowing the detection of UAVs at larger distances and/or under poor environmental conditions. In particular, this includes:
- Analysis and modelling of the expected false alarm and UAV target detection statistics.
- The implementation of a CFAR (Constant False Alarm Rate) algorithm with adaptive detection threshold matched to the expected false alarm statistics.
- The implementation of a radar based tracker, capable of initiating and maintaining track of a single UAV in adverse environmental conditions.
In a first step, the student will investigate several detection methods and will choose the best method for allowing UAV detection. The student will develop an algorithm for enhancing CFAR detection performance. Recent state of the art multi-target tracking algorithms based on Multi-Bernoulli Filters will be implemented and tested in terms of their abilities to initiate and maintain track of UAVs under varying distance and environmental conditions. The developed algorithms will ultimately form part of a vision/radar system.
- Literature review for radar based detection and tracking systems for UAVs
- Analysis of false alarm and UAV target detection statistics under various weather conditions.
- Implementation of an appropriate detection and single target tracking algorithm.
- Evaluate the algorithm performance in simulation.
- Literature review for radar based detection and tracking systems for UAVs - Analysis of false alarm and UAV target detection statistics under various weather conditions. - Implementation of an appropriate detection and single target tracking algorithm. - Evaluate the algorithm performance in simulation.
- Highly motivated and independent student.
- Interest in radar systems and probabilistic estimation algorithms.
- Excellent programming skills (in MATLAB and preferably also C++).
- Experience with working under the Robotic Operating System (ROS) is advantageous.
- Enrolled at ETH Zurich.
- Highly motivated and independent student. - Interest in radar systems and probabilistic estimation algorithms. - Excellent programming skills (in MATLAB and preferably also C++). - Experience with working under the Robotic Operating System (ROS) is advantageous. - Enrolled at ETH Zurich.
Amir Melzer (amir.melzer@mavt.ethz.ch), Martin Adams (martin.adams@mavt.ethz.ch)
Amir Melzer (amir.melzer@mavt.ethz.ch), Martin Adams (martin.adams@mavt.ethz.ch)