Fuzzy Modelling of Human Psycho-Physiological State and Fuzzy Adaptive Control of Automation in Human-Machine Interface

This research aims at proposing a new modelling and control framework that monitors the human operators' psychophysiological state in the human-machine interface to prevent performance breakdown. This research started with the exploration of new psychophysiological state assessment approaches to the adaptive modelling and control method for predicting human task performance and balancing the engagement of the human operator and the automatic system. The results of this research may also be further applied in developing advanced control mechanisms, investigating the origins of human compromised performance and identifying or even remedying operators' breakdown in the early stages of operation, at least.

A summary of the current human psychophysiological studies, previous human-machine interface simulation and existing biomarkers for human psychophysiological state assessment was provided for simulation experiment design of this research. The use of newly developed facial temperature biomarkers for assessing the human psychophysiological state and the task performance was investigated. The research continued by exploring the uncertainty of the human-machine interface system through the use of the complex fuzzy logic based offline modelling approach. A new type-2 fuzzy-based modelling approach was then proposed to assess the human operators' psychophysiological states in the real-time human-machine interface. This new modelling technique integrated state tracking and type-2 fuzzy sets for updating the rule base with a Bayesian process. Finally, this research included a new type-2 fuzzy logic-based control algorithm for balancing the human-machine interface systems via adjusting the engagement of the human operators according to their psychophysiological state and task performance. This innovative control approach combined the state estimation of the human operator with the type-2 fuzzy sets to maintain the balance between the task requirements (i.e. difficulty level) and the human operator feasible effort (i.e. psychophysiological states). In addition, the research revealed the impacts of multi-tasking and general fatigue on human operator's performance.