An Adaptive General Type-2 Fuzzy Logic Approach for Psychophysiological State Modeling in Real-Time Human-Machine Interfaces

In this article, a new type-2 fuzzy-based modeling approach is proposed to assess human operators' psychophysiological states for both safety and reliability of human-machine interface systems. Such a new modeling technique combines type-2 fuzzy sets with state tracking to update the rule base through a Bayesian process. These new configurations successfully lead to an adaptive, robust, and transparent computational framework that can be utilized to identify dynamic (i.e., real time) features without prior training. The proposed framework is validated on mental arithmetic cognitive real-time experiments with ten participants. It is found that the proposed framework outperforms other paradigms (i.e., an adaptive neuro-fuzzy inference system and an adaptive general type-2 fuzzy c-means modeling approach) in terms of disturbance rejection and learning capabilities. The proposed framework achieved the best performance compared to other models that have been presented in the related literature. Therefore, the new framework can be a promising development in human-machine interface systems. It can be further utilized to develop advanced control mechanisms, investigate the origins of human compromised task performance, and identify and remedy psychophysiological breakdown in the early stages.