基于对数正态误差函数的变步长盲均衡新算法

为了克服常数模算法(CMA)收敛速度慢,收敛后均方误差大的不足,定义了对数正态误差函数和新变步长,提出了基于对数正态误差函数的变步长盲均衡新算法。通过调整对数正态误差函数的均值与方差及变步长特性参数来加快该算法的收敛速度并减小收敛后的均方误差。用深海声道轴信道,对该算法的性能进行了仿真研究。结果表明:该算法收敛速度快、收敛后均方误差小,较好地克服了CMA的缺陷。

Novel Variable Step Size Blind Equalization Algorithm Based on Lognormal Error Function

When CMA (constant modulus algorithm) is used to equalize the underwater acoustic channels, there are two disadvantages: MSE (Mean Square Error) of the CMA is large and convergence is slow. For greatly reducing these two disadvantages, a new lognormal error function and new variable step size were defined, and a novel lognormal error function based variable step size blind equalization algorithm was proposed. In this proposed algorithm, the MSE could be reduced via regulating the mean-value and variance of lognormal error function, and the faster convergent speed could be obtained by changing the parameters charactering the variable step size. The computer simulation with the axial channel with deepwater acoustic track was carried out. The simulation results show that the proposed algorithm has faster convergent speed and smaller the MSE than the CMA. So it is especially fit for equalizing the underwater acoustic channel.