Adaptive neural Legendre orthogonal polynomial nonlinear channel equalization for chaos-based communications systems

Abstract

The performance of chaos-based communications systems is greatly affected by many sorts of nonlinear distortions. If nonlinear distortions in the channel can be removed, the performance of chaos-based communications systems can be improved. According to analysis of Volterra filter, a novel structure of neural network Legendre orthogonal polynomial equalizer is proposed based on the theory of chaotic signal reconstruction. Combining the characteristic of single layer neural network and structure of Legendre orthogonal polynomial, the equalizer is designed and realized after the analysis of a few parameter nonlinear filters, and adaptive algorithm is deduced using the normalized least mean square algorithm. To support the analysis, simulation results for nonlinear chaos-based communication channel are provided.