Force driven robot learning improves automated product disassembly success
Force-Driven Hook Disassembly via Constraint-Aware Object-Centric Reinforcement Learning
Robotics
Summary
Taking apart products at the end of their life is hard because some connections, like hooks, are tricky to undo without breaking. The authors designed a robot learning method that uses touch and force information, rather than just vision, to figure out how to pull hooks apart safely. Their robot learns from simple simulations and works well on real products it has never seen before. Adding a memory system helps the robot make smarter, less repetitive moves. This method showed better success and efficiency on various real disassembly tasks.
What this means in practice
- •For manufacturing automation teams: Enable robots to disassemble products with complex hook connections without customized models or prior demonstrations.
- •For robotics engineers: Improve robotic handling of objects constrained by hidden or deforming joints by integrating force-based learning with belief tracking.
Authors
Nadeen Elhadad, Takuya Kiyokawa, Kensuke Harada
Abstract
The growing volume of end-of-life products has intensified the need for automated disassembly, yet hook-type connections remain challengin: their orientations vary, they deform under load, and their geometry is hidden, so vision alone cannot identify a feasible non-destructive disassembly direction. This work proposes a constraint-aware reinforcement learning framework that infers feasible extraction directions online from contact forces. The policy is trained in a robot-agnostic, object-centric simulation on a simplified peg-in-hole pair and deployed zero-shot on a physical robot, where it transfers to real connections with unseen, multi-stage constraint structures. Because instantaneous contact forces are noisy and locally ambiguous, the policy is coupled with a particle filter belief that provides the policy with temporally integrated directional information and reduces redundant exploratory motion. The approach is validated on four real-world tasks, from peg-in-hole pairs to the inner propeller and front cover of an end-of-life air-conditioning unit. The full method achieves success rates of 100%, 100%, 90%, and 75% with the highest path efficiency on every task, and raises the success rate on the air-conditioning cover from 45% to 75% relative to a policy trained without the belief. These results indicate that force-driven reinforcement learning can enable disassembly without product-specific models or demonstrations.