Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces Critical-Manifold Guided RRT (CMG-RRT), a novel sampling-based motion planner tailored for rigid-body assembly under tight geometric constraints. By focusing on a critical manifold where valid motion paths exist, the method significantly enhances the efficiency of exploring configuration spaces in scenarios with minimal clearance. Empirical results show a remarkable 100% success rate in solving complex assembly tasks, including the previously unsolved Elk disentanglement puzzle, validating the approach's effectiveness.
Valid solution paths for tight assembly problems lie on a critical manifold, and CMG-RRT achieves a 100% success rate in solving complex assembly challenges.
Motion planning for rigid-body assembly poses a fundamental challenge in robotics due to tight geometric constraints. In such scenarios, feasible motions often require passing through (near-)zero clearance configurations in which the parts are tightly constrained by contact. In this work, we introduce Critical-Manifold Guided RRT (CMG-RRT), a sampling-based planner designed specifically for tight assembly problems. Our key observation is that in tight assemblies, valid solution paths lie on or near a critical manifold: the subset of configuration space consisting of poses with at least one contact point between parts. CMG-RRT guides exploration by adaptively biasing sampling toward neighborhoods of the critical manifold using a hierarchical subdivision of the configuration space. We prove that CMG-RRT is probabilistically complete under standard clearance assumptions. Empirical evaluation on challenging rotational assembly benchmarks demonstrates a 100% success rate across all tested instances, including, to the best of our knowledge, the first fully automatic solution of the Elk disentanglement puzzle. Our open source software is available through our project page: https://www.cgl.cs.tau.ac.il/projects/tight-assembly-planning.