Hierarchical fuzzy neural controller based on error iterative and approach

In this paper, a fuzzy neural networks based on hierarchical approach reasoning is proposed. The construction combining model is described by the fuzzy logic technology. The output of the antecedent part of the fuzzy logic is expressed as the input of the consequent part. The consequent part is a simple linear equation of the variables corresponding to the rule strength of the antecedent network and the output variables of the consequent network. So, the physical meaning of the proposed fuzzy neural network is clearer and its structure is simpler. We present a learning algorithm based on hierarchy error approach which utilizes a fuzzy logic function to aggregate the weight coefficients of the neural network, so the output can rapidly converge to the desired tolerable error range. Simulation results show the fuzzy neural network based on fuzzy hierarchy error approach have very good approach ability of for the complex functions through learning and training of the rule weight.