Real-time obstacle detection for Unmanned Aerial Vehicles
This thesis is done in collaboration between HIPPEROS and the embedded electronics group of the BEAMS department.
HIPPEROS is a spin-off of the BEAMS department developing real-time operating systems for embedded platforms. The HIPPEROS operating system exploits modern algorithms to better exploit available performances in such platforms to minimize costs, reduce energy consumption and provide safer autonomous systems.
Master Thesis description
The aim of this thesis is to enable real-time obstacle detection and navigation on an autonomous quadrotor. The quadrotor is a DJI Matrice-100, a developer-oriented quadrotor. The quadrotor is equipped with a LiDAR (Light Detection And Ranging) system (similar to what is used on Google’s self-driving car), which can measure distances to the nearest obstacle. The quadrotor and the LiDAR are controlled by a microprocessor system integration the HIPPEROS OS.
This project aims at demonstrating the performances of the HIPPEROS real-time OS with such a platform. To this end, the student will have to do the following:
- Design an obstacle detection algorithm using the LiDAR mounted on the quadrotor. The obstacle detection will have to satisfy real-time constraints, which is possible thanks to the HIPPEROS real-time OS;
- Design a navigation algorithm that is able to avoid the detected obstacles. The movement characteristics of the quadrotor will depend on the obstacle detection speed determined in the first point;
- Show the flight data on a host machine (e.g. through WiFi), as well as the different detected obstacles.
Supervisor(s): Prof. François Quitin, Olivier Desenfans (HIPPEROS)
Note: it is highly recommended to combine this thesis with an internship at HIPPEROS
Information : François Quitin (fquitin [at] ulb [dot] ac [dot] be) Tel : 02-650-2829, BEAMS Department
Students : ELEC, INFO, EM