Direct-Inverse Modeling Control Based on Interval Type-2 Fuzzy Neural Network

This paper presents direct-inverse modeling control method based on interval type-2 fuzzy neural networks. This control method includes two phases, i. e., structure identification and parameters learning. In structure identification phase, hierarchical fuzzy clustering method is used to identify the initial structure of interval type-2 fuzzy neural network at first. Then, the uncertain parameters of Gauss membership functions of interval type-2 fuzzy sets are decided. In parameters learning phases, BP algorithm of interval type-2 fuzzy neural networks is adopted to adjust the free parameters of precondition and consequence. At last, inverse model of controlled plant is identified in the off-line manner as the controller. The simulation experiment of a single-input and single-output nonlinear system shows that this proposed control method is effective.