Lightweight Design of Robotic Arm Based on Multimodal Large Language Models

With the rapid development of society, the application of robotic arms is becoming increasingly widespread. Lightweighting of robotic arms can not only increase flexibility and operability, but also reduce costs and energy. In recent years, Large Language Models (LLMs) have witnessed rapid development. Applying LLMs to the lightweight design of robotic arms can enhance design efficiency and reduce design difficulties. In this article, a method that directly utilizes prompt engineering to guide a multimodal LLM (MLLM) in automatically generating lightweight design schemes for the upper arm of a SCARA robot is proposed. Firstly, static simulation of the upper arm is conducted using ANYSY via Python. Secondly, utilizing the MLLM and prompt engineering, automatic generation of lightweight design schemes for drilling is achieved, with corresponding Python code generated, based on input stress diagram. Finally, the geometry model is reconstructed and subjected to static simulation. The results showed that the lightweight design scheme generated by MLLM achieved the optimization goal, yielding results comparable to those of existing works. The new method using MLLM for the lightweight design of robot arms is able to effectively improve design efficiency and reducing design difficulty, with broad application prospects.