Behavioral modeling by Volterra series and time-delay neural network approach

Abstract

Solving the non-linear distortion problems in wireless communications is often based on developing the behavioral models of non-linear components. In this paper, a non-linear Volterra model up to third order is developed by using an artificial neural network (ANN) approach. The Volterra kernels are derived from the parameters of a feed-forward time delay neural network with a suitable activation function. The Volterra model is implemented to represent the behavior of a low noise amplifier for LTE receiver.