A Fuzzy Neural Network Model and Its Hardware Implementation

In this paper, a fuzzy classifier based on a four-layered feedforward neural network model is proposed. This connectionist fuzzy classifier, called CFC, realizes the “weighted Euclidean distance” fuzzy classification concept in a massively parallel manner to recognize input patterns. CFC employs a hybrid supervised/unsupervised learning scheme to organize referenced pattern vectors. This scheme not only overcomes the major drawbacks of the multilayer perceptron models with the back-propagation algorithm: local minimal problem and long training time, but also avoids the disadvantage of the huge storage space requirement of the probabilistic neural network (PNN). According to experimental results, CFC shows better accuracy on speech recognition than several existing methods, even in a noisy environment. Moreover, it has a higher stability on recognition rates for different environmental conditions. To implement CFC with digital hardware, a massively parallel hardware architecture is developed. In this architecture, the techniques of bus-oriented multiprocessor, systolic processor structure, and pipelining are applied to provide a low-cost, high-performance fuzzy classification scheme. The systolic VLSI processor for membership grade computation has been modeled with the VHDL language and synthesized by a high-level synthesis tool. Fuzzy neural network, fuzzy classification, hierarchical clustering, speech recognition, massively parallel hardware, systolic VLSI processor, VHDL language, high-level synthesis.