A logic-assisted assembly support system to promote the inclusion of neurodiverse operators in manufacturing

Inclusive manufacturing fosters social sustainability and equal opportunities for diverse cognitive profiles, achievable through Industry 4.0/5.0 technologies. A logic-assisted assembly support (LoAS) system to guide neurodiverse operators with logic-related assembly challenges is proposed in this research. The system uses a deep learning paradigm for intelligent object recognition to verify the correct completion of critical steps and to identify assembly errors and incorrect sequences. The operational logic is modelled through a dynamic flowchart automatically generated by a large language model (LLM) based on feedback from the vision system. This flowchart shows each step, along with the tools and components required, and is updated in real time to inform operators of their next steps and current location within the process. The system also includes a step-by-step verification mechanism with detailed completion diagrams to ensure accuracy. Interactive instructions provide personalised step-by-step guidance, visual identification and immediate feedback to correct errors and guide the operator through the assembly sequence. The efficiency and effectiveness of LoAS have been verified at a proof-of-concept lab scale on an industrial workpiece. The results suggest the potential of those tools to foster inclusiveness in manufacturing environments.