CoViLLM: An Adaptive Human-Robot Collaborative Assembly Framework Using Large Language Models

TL;DR

CoViLLM introduces an adaptive human-robot assembly framework that leverages large language models, depth sensing, and human classification to enable flexible assembly of customized and unseen products.

Contribution

The paper presents a novel framework combining LLMs with perception and human classification to autonomously plan and execute assembly tasks for new product variants.

Findings

Enables assembly of unseen products with high flexibility

Uses LLMs for natural language task planning

Validated on NIST Assembly Task Board with positive results

Abstract

With increasing demand for mass customization, traditional manufacturing robots that rely on rule-based operations lack the flexibility to accommodate customized or new product variants. Human-Robot Collaboration has demonstrated potential to improve system adaptability by leveraging human versatility and decision-making capabilities. However, existing Human-Robot Collaborative frameworks typically depend on predefined perception-manipulation pipelines, limiting their ability to autonomously generate task plans for new product assembly. In this work, we propose CoViLLM, an adaptive human-robot collaborative assembly framework that supports the assembly of customized and previously unseen products. CoViLLM combines depth-camera-based localization for object position estimation, human operator classification for identifying new components, and a Large Language Model for assembly task planning based on natural language instructions. The framework is validated on the NIST Assembly Task Board for known, customized, and new product cases. Experimental results show that the proposed framework enables flexible collaborative assembly by extending Human-Robot Collaboration beyond predefined product and task settings.