A hierarchical fuzzy approach for adaptation of pre-given parameters in an interval type-2 TSK fuzzy neural structure

In self-evolving type-2 fuzzy neural structures, there are several pre-given parameters that are conventionally defined before the runtime by using trial-and-error. This approach is very time-consuming and does not guarantee that the selected values are the most appropriate ones for ensuring high convergence speed. To overcome these drawbacks, here a hierarchical fuzzy controller is proposed. The proposed hierarchical controller helps to increase precision since it dynamically adjusts pre-given parameters online by considering the error changes. Moreover, the proposed structure helps to reduce complexity and avoid “curse of dimensionality” which is a common phenomenon when the number of input variables to the fuzzy system is large. Hence, this structure is suitable for type-2 fuzzy neural systems which usually have several pre-given parameters to be adjusted. The proposed hierarchical fuzzy controller is applied to an interval type-2 TSK fuzzy neural network and the performance is investigated by comparing the results with trial-and-error approach in two different applications of identification and control. The simulation results indicate that the proposed method can effectively cover the drawbacks of trial-and-error approach while it enhances the precision of the system.