Keywords: Ergodic Control, Surface Finishing, Robots, Learning, Collaboration
Abstract: Surface finishing is a critical yet challenging task in manufacturing, often requiring significant manual effort and expertise. This paper presents a
novel approach to automate surface finishing using collaborative robots that learn from human demonstrations. By leveraging user-friendly programming methods and ergodic control techniques, our system enables real-time task execution and interaction. The proposed method maps human demonstrations onto a two-dimensional parametric space, learns the task description using a Gaussian mixture model, and employs an ergodic controller to guide the robot. Evaluation on a 7-DoF torque controlled robot demonstrates the effectiveness of our approach
in generating accurate task descriptions and facilitating user interaction during execution.
Submission Number: 8
Loading