Development of an AI-based informational assistance system for assembly accounting for dyslexia

The manufacturing sector faces significant challenges in promoting inclusivity, particularly for neurodiverse individuals who may encounter difficulties with traditional text-based assembly instructions. This research addresses such challenges by developing a worker-assistance framework designed to improve both operational efficiency and inclusivity within assembly tasks. The proposed AI-based system integrates an intelligent vision system for real-time task monitoring, an intelligent instruction generation module for producing personalised, context-specific guidance, and an instruction delivery module that provides hands-free, voice-guided instructions. In a laboratory-scale case study involving the assembly of a centrifugal pump, the system was tested with simulated cognitive challenges, such as dyslexia-like text distortions. The results indicate that the AI-driven system could significantly decrease assembly errors and task completion time compared to traditional human supervision, while providing support tailored to the needs of neurodiverse operators. These findings suggest the system potential to prevent common errors and improve accessibility for operators with cognitive variations. Future developments may enhance system flexibility for different assembly workflows, introduce techniques to evaluate the cognitive effort, and extend its implementation to a more diverse neurodivergent workforce to strengthen inclusivity in manufacturing environments.