Augmented Reality for Maintenance Tasks with ChatGPT for Automated Text-to-Action
share
Advancements in sensor technology, artificial intelligence (AI), and augmented reality (AR) have unlocked opportunities across various domains. AR and large language models like GPT have witnessed substantial progress and are increasingly being employed in diverse fields. One such promising application is in operations and maintenance (O M). O M tasks often involve complex procedures and sequences that can be challenging to memorize and execute correctly, particularly for novices or under high-stress situations. By marrying the advantages of superimposing virtual objects onto the physical world, and generating human-like text using GPT, we can revolutionize O M operations. This study introduces a system that combines AR, Optical Character Recognition (OCR), and the GPT language model to optimize user performance while offering trustworthy interactions and alleviating workload in O M tasks. This system provides an interactive virtual environment controlled by the Unity game engine, facilitating a seamless interaction between virtual and physical realities. A case study (N=15) is conducted to illustrate the findings and answer the research questions. The results indicate that users can complete similarly challenging tasks in less time using our proposed AR and AI system. Moreover, the collected data also suggests a reduction in cognitive load and an increase in trust when executing the same operations using the AR and AI system.
READ FULL TEXT