SELF-VLA: A Skill Enhanced Agentic Vision-Language-Action Framework for Contact-Rich Disassembly

2026-03-10

AI Summary

Existing robotic disassembly approaches rely heavily on task-specific data and struggle to handle the high variability, strong contact interactions, and long-horizon sequential operations inherent in end-of-life electronics disassembly, resulting in poor generalization and deployment challenges. This work proposes a vision–language–action (VLA) agent framework that integrates explicit disassembly skills, marking the first effort to embed structured disassembly knowledge into a VLA model. By incorporating modular skill design, contact-aware control, and sequential decision-making mechanisms, the approach overcomes the limitations of end-to-end VLA models in complex industrial disassembly tasks. Experiments demonstrate that the proposed method significantly outperforms state-of-the-art VLA models on two contact-intensive disassembly tasks, achieving substantial improvements in both generalization capability and practical applicability.

Technology Category

• Intelligent Robots: Manipulation
• Computer Vision: Language and Vision
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: Architectures and workflows that use LLMs for crowd work

Abstract

Disassembly automation has long been pursued to address the growing demand for efficient and proper recovery of valuable components from the end-of-life (EoL) electronic products. Existing approaches have demonstrated promising and regimented performance by decomposing the disassembly process into different subtasks. However, each subtask typically requires extensive data preparation, model training, and system management. Moreover, these approaches are often task- and component-specific, making them poorly suited to handle the variability and uncertainty of EoL products and limiting their generalization capabilities. All these factors restrict the practical deployment of current robotic disassembly systems and leave them highly reliant on human labor. With the recent development of foundation models in robotics, vision-language-action (VLA) models have shown impressive performance on standard robotic manipulation tasks, but their applicability to complex, contact-rich, and long-horizon industrial practices like disassembly, which requires sequential and precise manipulation, remains limited. To address this challenge, we propose SELF-VLA, an agentic VLA framework that integrates explicit disassembly skills. Experimental studies demonstrate that our framework significantly outperforms current state-of-the-art end-to-end VLA models on two contact-rich disassembly tasks. The video illustration can be found via https://zh.engr.tamu.edu/wp-content/uploads/sites/310/2026/03/IROS-VLA-Video.mp4.

Problem

• disassembly automation
• contact-rich manipulation
• vision-language-action
• generalization
• end-of-life electronics

Innovation

• Vision-Language-Action
• Skill Integration
• Contact-Rich Disassembly
• Agentic Framework
• Foundation Models