Reproduce “Harnessing with Twisting: Single-Arm Deformable Linear Object Manipulation for Industrial Harnessing Task” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Xiang Zhang, Hsien-Chung Lin, Yu Zhao, Masayoshi Tomizuka
Researchers at UC Berkeley and FANUC developed a single-robot arm pipeline that automates industrial wire harnessing by generating and controlling wire tension through a novel twisting motion. This system, which integrates a Koopman operator-based Model Predictive Control, achieved a 100% success rate on single-wire tasks and robust performance on multi-wire tasks on the NIST Assembly Task Board 4, outperforming conventional methods.
Key insights
- Actively twisting the wire with a single robot arm provides a robust mechanism to generate and maintain sufficient tension, which is crucial for reliably inserting DLOs into tight industrial clamps.
- The Koopman operator effectively linearizes the complex, nonlinear dynamics of wire manipulation, enabling stable and accurate tension tracking through a standard MPC framework.
- Integrating a specialized gripper, task-specific motion planning, and dynamic fix-point updates allows for a comprehensive and robust single-arm solution for multi-stage industrial harnessing tasks.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.