Prediction and separation of synchronous-networking frequency hopping signals based on RBF neural network

Abstract

In this paper, a novel method of prediction and separation of Synchronous-networking frequency hopping (FH) signals is proposed. By choosing suitable chaotic sequences which have pseudo-orthogonal property and global phase diagram with single mapping path, a Radial Basis Function (RBF) neural network can be trained for the prediction and separation of multi FH signals. First, the performance of chaotic sequence is analyzed, based on which the inputs and outputs for the training of RBF neural network are selected. Then the algorithm of Orthogonal Least Square (OLS) is applied for the predictor and separator. Simulation results show that the algorithm can effectively predict and separate the synchronous-networking FH signals.