SAE Aero Design · Ongoing project
Autolander
Embedded vision system for precision landing of an autonomous cargo aircraft in the AeroDesign competition.
Context
Autolander is part of the advanced mission of the SAE Aero Design competition, where the aircraft must take off, deliver or recover a payload, then return to base within a limited time. The project brings together the Club d'intelligence artificielle from computer science and Avion Cargo ULaval from mechanical engineering around one embedded system.
My role
I work on embedded vision software and system integration. I structure the Python code, develop ArUco detection with OpenCV, estimate distance and orientation relative to the marker, prepare MAVLink communication with ArduPilot, and contribute to camera-stream visualization during testing in coordination with mechanical and aeronautical constraints.
Features
- ArUco marker detection from an onboard camera
- Interdisciplinary collaboration with an aeronautics team
- 3D pose estimation with OpenCV
- Guidance toward a precision landing zone
- MAVLink communication with the ArduPilot controller
- Real-time camera stream visualization
- Simulation and flight-test preparation
Technologies
- Python
- OpenCV
- ArUco
- Raspberry Pi
- ArduPilot
- MAVLink
- Video streaming
- Gazebo
- Poetry