{"ID":2876580,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2508.21272","arxiv_id":"2508.21272","title":"Learning to Assemble the Soma Cube with Legal-Action Masked DQN and Safe ZYZ Regrasp on a Doosan M0609","abstract":"This paper presents the first comprehensive application of legal-action masked Deep Q-Networks with safe ZYZ regrasp strategies to an underactuated gripper-equipped 6-DOF collaborative robot for autonomous Soma cube assembly learning. Our approach represents the first systematic integration of constraint-aware reinforcement learning with singularity-safe motion planning on a Doosan M0609 collaborative robot. We address critical challenges in robotic manipulation: combinatorial action space explosion, unsafe motion planning, and systematic assembly strategy learning. Our system integrates a legal-action masked DQN with hierarchical architecture that decomposes Q-function estimation into orientation and position components, reducing computational complexity from $O(3,132)$ to $O(116) + O(27)$ while maintaining solution completeness. The robot-friendly reward function encourages ground-first, vertically accessible assembly sequences aligned with manipulation constraints. Curriculum learning across three progressive difficulty levels (2-piece, 3-piece, 7-piece) achieves remarkable training efficiency: 100\\% success rate for Level 1 within 500 episodes, 92.9\\% for Level 2, and 39.9\\% for Level 3 over 105,300 total training episodes.","short_abstract":"This paper presents the first comprehensive application of legal-action masked Deep Q-Networks with safe ZYZ regrasp strategies to an underactuated gripper-equipped 6-DOF collaborative robot for autonomous Soma cube assembly learning. Our approach represents the first systematic integration of constraint-aware reinforc...","url_abs":"https://arxiv.org/abs/2508.21272","url_pdf":"https://arxiv.org/pdf/2508.21272v1","authors":"[\"Jaehong Oh\",\"Seungjun Jung\",\"Sawoong Kim\"]","published":"2025-08-29T00:27:03Z","proceeding":"cs.RO","tasks":"[\"cs.RO\",\"stat.CO\"]","methods":"[\"Reinforcement Learning\"]","has_code":false}