This project was completed for Boston Unviversity EK505 as a Final Project.
The purpose of this project was to design a software system that would allow a warehouse robot with an overhead rgb camera to detect and drive around obstacles. We explored using Meta AI's Segment Anything Model and other segmentation algorithms to segment our live camera feed, from which we detect obstacles. We then use a CLF-CBF formulation to drive the robot around the obstacles to the goal point.