A neuro-fuzzy hybrid scheme for design and simulation of human machine systems

Abstract

In this article, a new neuro-fuzzy hybrid approach to human workplace design and simulation is proposed. Problems related to human workplace design such as human-machine modeling, measurement and analysis, workplace layout design and planning, workplace evaluation and simulation are discussed in detail. The complex human-machine interactions in workplace design are described with human and workstation parameters within a comprehensive human-machine system model. Based on this model, procedures and algorithms for workplace design, ergonomic evaluation, and optimization are presented in an integrated framework. With a combination of individual neural and fuzzy techniques, the neuro-fuzzy hybrid scheme implements fuzzy if-then rules block for workplace design and evaluation by trainable neural network architectures. For training and test purposes, simulated assembly tasks are carried out on a self-built multiadjustable laboratory workstation with a flexible PEAK Motus motion measurement and analysis system. The trained fuzzy neural networks are capable of predicting the operator's posture and joint angles of motion associated with a range of workstation configurations. They can also be used for design/layout and adjustment of manual assembly workstations. The developed system provides a unified, intelligent computational framework for human-machine system design and simulation. In the end, case studies for workplace design and simulation are presented to validate and illustrate the developed neuro-fuzzy design scheme and system.