混沌通信系统中非线性信道的自适应神经Legendre正交多项式均衡

针对混沌通信系统中非线性信道干扰问题，基于混沌信号重构理论和Legendre正交多项式结构，提出了一种自适应神经Legendre正交多项式信道均衡器，并给出相应的归一化最小均方算法. 仿真研究表明：所提出的自适应神经Legendre正交多项式信道均衡器能有效地消除线性和非线性信道干扰，均衡器输出信号能反映出混沌信号的特性，具有良好的抗干扰性能.该均衡器的结构简单，权系数参数较少，收敛稳定性较好. 关键词： Legendre 正交多项式 信道均衡 混沌吸引子 神经网络

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

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.