Design of Photovoltaic Cleaning Robots Assisted by Large Language Models

With the booming development of the photovoltaic (PV) industry, the cleaning of PV panels has become increasingly crucial. This study, framed under the double diamond model, utilized the large language model (LLM) GPT-4 to assist in designing a PV cleaning robot. Initially, the study involved in-depth market analysis and field investigations to identify consumer needs and design requirements. Subsequently, the Kano model was applied to effectively categorize and assess consumer needs, successfully pinpointing key design directions. GPT-4 proposed numerous innovative design ideas. We created the selected prototypes to test and chose the most promising one ultimately. The results indicate that the overall performance of the scheme met expectations, demonstrating the potential of LLMs in complex product design.