The goal of this project was to create an autonomous parking system for 1/10th scale cars. When the automated parking system is run the car will initiate a 3 subsystem process:
- The perception subsystem uses a 3D point-cloud dataset from a ZED stereo camera to find the center of the closest available parking space, as well as any potential obstacles
- The path-planning subsystem uses this parking spot centroid as an end point to chart a Bezier curve path from the car to the spot
- The control subsystem converts this planned path into discrete direction commands to send to the VESC motor controller to physically move the vehicle.
Should the vehicle detect an obstacle in the way or loses sight of the parking space the vehicle safely comes to a stop.
Team Members: Marko Buha, Quinn Gauder, Ryan Gowland, Beichen Huang, Mohammad Nasser
Advisors: Stefan Ramdhan, Dr. Victor Bandur, Dr. Vera Pantelic, Dr. Alan Wassyng
Partner: Physical vehicle provided by the McMaster Centre for Software Certification (McSCert)