The Fuzzy Modeling Method’s Research Of Operator Functional State

In the safety-critical complex human-machine systems, the operator requires to accomplish a certain task gether with the automation systems. Among them, the operator’s performance has become a key factor in determining the performance of the system.In order to keep the systems running safely, it can monitor and analyze the operator functional state (OFS) efficiently. And we can adjust the task load according to the performance of the operator to prevent the failure of the systems and the potential accidents.Based on the operator’s electrophysiological signals and performance data, this paper use the fuzzy modeling methods for assess the OFS quantitatively. Firstly, through the effective data preprocessing, we choose the input and output variables of the models. Secondly, we use the Wang-Mendel (WM) fuzzy modeling method to build the static OFS moddels. We use three different rule-extraction WM methods (Method1, Method2, Method3) to get the rules and compare the fitting results. The comparison results indicate that we can get more excellent performance static OFS models using the Method3. And we use the Recurrent Neuro-Fuzzy Network (RNF) and the WM methods to get the RNF-WM method. Then we use the RNF-WM method to build the dynamic OFS models. This dynamic model can use of the effective data beter so that the model can be updated in real time and adapted to the changes of the OFS. In addition, this paper use the particle swarm optimization (PSO) optimize the parameters of the static and dynamic OFS models.So we can improve the accuracy of the OFS static and dynamic models. As a preliminary research for the OFS modeling, this paper provides a model support for the future application of the adaptive human-machine automation system.