USING SELF-CONSTRUCTING RECURRENT FUZZY NEURAL NETWORK AS A DIGITAL CHANNEL EQUALIZER

摘要

In this paper the design of a self-constructing recurrent fuzzy neural network (SCRFNN)-based digital channel equalizer is proposed. It is found that a digital channel equalizer based on SCRFNN can recover channel distortions effectively. We compare the performance of SCRFNN with adaptive-based-network fuzzy inference system (ANFIS) and the Bayesian equalizers in complex-valued linear channels. Our simulations show that the performance of SCRFNN is close to the Bayesian optimal solution. Furthermore, the hardware requirement of the trained SCRFNN equalizer is relatively lower than the other two structures.