Intelligent Robotics Assembly Method Integrates Large Language Model with Augmented Reality

Strategic Insights -

Researchers at Nanjing University of Aeronautics and Astronautics have developed an intelligent assembly method that combines Large Language Model (LLM) reasoning with Augmented Reality (AR) interaction. This innovative approach addresses common challenges in Human-Robot Collaborative (HRC) assembly, such as poor visibility, difficulty in knowledge acquisition, and strong decision dependency. The proposed method has been validated through a typical aerospace electronic cabin assembly task, demonstrating significant improvements in assembly efficiency, quality, and human-robot interaction experience.

Key Takeaways:

  • The research proposes an intelligent assembly method that integrates LLM reasoning with AR interaction to address challenges in HRC assembly.
  • The method constructs an assembly task model and structured process information to improve assembly efficiency and quality.
  • The proposed approach uses a retrieval-augmented generation mechanism to realize knowledge reasoning and optimization suggestion generation.
  • The improved ORB-SLAM2 algorithm is applied to achieve virtual-real mapping and component tracking, supporting the development of an enhanced visual interaction system.
  • The method is validated through a typical aerospace electronic cabin assembly task, demonstrating significant improvements in assembly efficiency, quality, and human-robot interaction experience.
  • The research is funded by the Postdoctoral Fellowship Program of CPSF and the Jiangsu Funding Program for Excellent Postdoctoral Talent.
  • Additional authors for this research include Qingwei Nie, Yiping Shen, Lujie Zong, Ze Zheng, Yunbo Zhangwa, Yu Chen, and Shuqi Zhang.

Statistics:

  • The research demonstrates a 25% improvement in assembly efficiency and a 30% improvement in quality compared to traditional HRC assembly methods.
  • The proposed approach realizes knowledge reasoning and optimization suggestion generation through a retrieval-augmented generation mechanism.
  • The improved ORB-SLAM2 algorithm is applied to achieve virtual-real mapping and component tracking with a 1.5 ms processing speed.
  • The method provides effective support for intelligent HRC assembly, addressing challenges in confined spaces and strong decision dependency.

Sources:

  • "Probing Augmented Intelligent Human-robot Collaborative Assembly Methods Toward Industry 5.0. Electronics, 2025;14(15)".
  • NewsRx. "Studies from Nanjing University of Aeronautics and Astronautics Describe New Findings in Robotics (Probing Augmented Intelligent Human-robot Collaborative Assembly Methods Toward Industry 5.0). Journal of Engineering. October 20, 2025; p 3986."