Learning to Assemble the Soma Cube with Legal-Action Masked DQN and Safe ZYZ Regrasp on a Doosan M0609

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Oh, Jaehong, Jung, Seungjun, Kim, Sawoong
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912558801223680
author Oh, Jaehong
Jung, Seungjun
Kim, Sawoong
author_facet Oh, Jaehong
Jung, Seungjun
Kim, Sawoong
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2508_21272
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Learning to Assemble the Soma Cube with Legal-Action Masked DQN and Safe ZYZ Regrasp on a Doosan M0609
Oh, Jaehong
Jung, Seungjun
Kim, Sawoong
Robotics
Computation
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.
title Learning to Assemble the Soma Cube with Legal-Action Masked DQN and Safe ZYZ Regrasp on a Doosan M0609
topic Robotics
Computation
url https://arxiv.org/abs/2508.21272