Learning-Guided Force-Feedback Model Predictive Control with Obstacle Avoidance for Robotic Deburring

2026-04-07

AI Summary

This work addresses the limitations of traditional model predictive control (MPC) in contact-intensive industrial tasks such as deburring, where simultaneous real-time force feedback, stable force regulation, and collision avoidance—particularly during precise insertion and circular motions in confined spaces—remain challenging. The authors propose the first integration of a diffusion model as a motion prior within a force-feedback MPC framework. This approach leverages the diffusion model to generate robust initial trajectories and enhance cross-task generalization, while MPC ensures accurate normal force tracking, feasible joint torque execution, and real-time collision avoidance during operation. Experiments on a real torque-controlled robotic arm demonstrate that the method reliably accomplishes complex tasks—including high-precision tool insertion and obstacle-constrained circular deburring—even in hard-to-reach configurations, thereby overcoming the performance bottlenecks of conventional MPC under strong contact interactions and tight geometric constraints.

Abstract

Model Predictive Control (MPC) is widely used for torque-controlled robots, but classical formulations often neglect real-time force feedback and struggle with contact-rich industrial tasks under collision constraints. Deburring in particular requires precise tool insertion, stable force regulation, and collision-free circular motions in challenging configurations, which exceeds the capability of standard MPC pipelines. We propose a framework that integrates force-feedback MPC with diffusion-based motion priors to address these challenges. The diffusion model serves as a memory of motion strategies, providing robust initialization and adaptation across multiple task instances, while MPC ensures safe execution with explicit force tracking, torque feasibility, and collision avoidance. We validate our approach on a torque-controlled manipulator performing industrial deburring tasks. Experiments demonstrate reliable tool insertion, accurate normal force tracking, and circular deburring motions even in hard-to-reach configurations and under obstacle constraints. To our knowledge, this is the first integration of diffusion motion priors with force-feedback MPC for collision-aware, contact-rich industrial tasks.

Problem

• Model Predictive Control
• Force Feedback
• Obstacle Avoidance
• Robotic Deburring
• Contact-Rich Tasks

Innovation

• diffusion-based motion priors
• force-feedback MPC
• obstacle avoidance
• contact-rich manipulation
• robotic deburring