The implementation of fuzzy systems, neural networks and fuzzy neural networks using FPGAs

This paper describes the hardware implementations of fuzzy systems, neural networks and fuzzy neural networks (FNNs) using Xilinx Field Programmable Gate Arrays (FPGAs). The validity of these approaches is demonstrated by their application to a non-linear function approximation problem. The various elements of each system are discussed and implemented in hardware. The architectures were also implemented in software using the MATLAB neural network toolbox. The results are analysed in terms of an accuracy performance index and in the dimensions of the hardware required. (C) 1998 Elsevier Science-Inc. All rights reserved.