Computer-Vision-Driven Artificial Potential Function Guidance and Adaptive Control for Spacecraft Proximity Operations

This research addresses the growing issue of space debris by developing advanced computer vision, guidance, and control techniques for autonomous docking in proximity operations. Specifically, this work develops these technologies to present an experiment where a chaser platform autonomously docks with a cooperative spinning target while avoiding an uncooperative obstacle. A stereovision system using ArUco markers tracks the target’s pose in real-time, while an unscented Kalman filter processes the data. The obstacle is detected through bounding box manipulation and stereo disparity principles. A novel artificial potential function guidance law, herein adapted for spinning targets, calculates a collision-free trajectory, which is tracked using a real-time adaptive control law. Experimental validation at Carleton University’s Spacecraft Proximity Operations Testbed confirms the effectiveness of the proposed system.