水声信道变步长自适应均衡算法

Variable Step-Size Adaptive Equalization Algorithm for Underwater Acoustic Channel

针对定步长自适应均衡算法收敛速度与稳态性能之间的矛盾,提出一种适用于自适应均衡器的变步长仿射投影算法。该算法考虑到传统仿射投影算法中约束条件迫使自适应滤波器补偿干扰信号的影响,当步长较大且干扰信号存在时,引起稳态性能下降的问题,通过建立一个新的约束条件,并利用先验误差向量的定义,得到一种适用于仿射投影均衡算法的步长调节方法。理论分析和计算机仿真结果表明:所提算法与定步长仿射投影算法和变步长最小均方算法相比,具有更快的收敛速度和更低的失调率和误码率。

The performance of constant step-size adaptive equalization algorithm is governed by a step size parameter.The choice of this parameter re⁃flects the tradeoff between fast convergence and poor steady state misalignment.To deal with this issue,a novel variable step size affine pro⁃jection algorithm is proposed.The constraints of the traditional affine projection algorithm force the adaptive filter to compensate for the ef⁃fect of noise signal.As long as the noise signal is not zero,the updated coefficient vector cannot be equal to the real parameters of the un⁃known system.In order to overcome this effect,the proposed algorithm obtained an ideal constraint function by forcing the posterior error to cancel negative effect of disturbance signal,Then using this function,a step size control approach for affine projection algorithm is provid⁃ed.The theoretical analysis and computer simulation results verify that the proposed algorithm can retain fast convergence rate and low fi⁃nal misalignment in underwater acoustic channel equalization.