Regulation of Adaptive Human-Machine System Based on Fuzzy Inference Petri Net

Abstract

In Human-Machine (HM) systems, human operators and machines cooperate to achieve a system goal. Accordingly, the operators' task load should be dynamically adjusted in accordance with their physiological and psychological status, so-called operator functional state (OFS). In this sense, the HM system behavior becomes hybrid. Petri net and its extensions have been widely used in hybrid system modeling and control. In this paper, we use a fuzzy inference Petri net to model the HM system and then realize adaptive task allocation between human operator and machine. The model can predict the “performance” of the operator, based on which the workload can be reallocated (i.e., increased or reduced somehow) between the operator and machine in an adaptive fashion the tasks. The simulation results demonstrate that the approach of fuzzy inference Petri nets proposed is an effective way in modeling and control of HM hybrid systems.