From insight to action: Embodied multi-agent system integrating vision language model for digital twin-assisted human-robot collaborative assembly

Key Points

• The integration of a vision language model significantly improves the performance of multi-agent systems in tasks related to collaborative assembly.
• This approach demonstrated a notable increase in assembly efficiency, with preliminary tests showing a 30% reduction in completion time compared to traditional methods.
• Assessment using digital twin technology facilitates real-time interaction and planning, enhancing human-robot collaboration during assembly processes.
• The findings support the potential for advanced multi-agent systems in manufacturing, calling for broader applications in real-world settings.