用安全抓取与约束强化学习，让机械臂自主拼装索马立方体。
Learning to Assemble the Soma Cube with Legal-Action Masked DQN and Safe ZYZ Regrasp on a Doosan M0609
- 用分层DQN分解动作空间，降低计算复杂度至116+27
- 三阶段训练达成92.9%成功率，7块组合达39.9%
- 适合研究机器人自主装配与安全运动规划的开发者
本文首次将合法动作掩码深度Q网络与安全ZYZ重抓策略应用于配备欠驱动夹爪的6自由度协作机器人，实现索马立方体的自主组装学习。系统首次整合了约束感知强化学习与奇异性安全运动规划。针对动作空间爆炸、运动规划不安全及系统性装配策略学习等挑战，采用分层架构将Q函数估计分解为方位与位置两部分，计算复杂度由$O(3,132)$降至$O(116) + O(27)$，同时保证解的完备性。设计符合机器人操作约束的奖励函数，引导地面优先、垂直可达的组装序列。通过三个渐进难度层级（2件、3件、7件）的课程学习，在总计105,300次训练中，一级任务500轮内达成100%成功率，二级任务92.9%，三级任务39.9%。
原文摘要 · Abstract (English)
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.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。