Simulation of Gaussian potential function network for adaptive equalization

This study consists of simulation of Gaussian Potential Function Network (GPFN) [1] and application to filtering problem called intersymbol interference (ISI) equalization which is conventionally handled with digital equalizers. A GPFN is a Radial Basis Function (RBF) like net. It has three layers; one input layer, one hidden layer, which includes the nodes with gaussian activation function, and the last is the output layer. Its distinct feature comes from its ability of automatic recruitment of hidden layer nodes. The filtering of signals, which are corrupted with additive white gaussian noise (AWGN) to represent ISI, performed with this simulation showed that this network is a suitable tool for adaptive equalization.