水声信道混合型常数模盲均衡新算法

为了克服常数模算法(constantmodulusalgorithm,CMA)收敛速度慢、符号误差常数模算法(signerrorcon stantmodulusalgorithm,SECMA)收敛后剩余干扰大的缺点,利用常数模代价函数和符号误差常数模代价函数定义了一个混合常数模新代价函数,并从理论上分析了该代价函数的特性,构造了具有常数模代价函数梯度矢量与符号误差常数模代价函数梯度矢量联合更新的混合常数模新算法(mixedconstantmodulusalgorithm,MCMA)。调整该算法中两个梯度矢量的比例系数,可提高该算法收敛速度、减少收敛后的均方误差。通过对水声信道均衡的计算机仿真,证明了该算法的有效性。

Novel mixed constant modulus blind equalizer for underwater acoustic channel

For greatly improving the convergence rate of constant modulus algorithm(CMA) and reducing the mean square error of sign error constant modulus algorithm(SECMA), a novel mixed constant modulus cost function is defined using constant modulus cost function and sign error constant modulus cost function, and its feature is analyzed in theory. Mixed constant modulus algorithm(MCMA), which can be updated by gradient vector of constant modulus cost function and that of sign error constant modulus cost function, is developed. In the MCMA, adjusting the scale coefficients of two gradient vectors can improve the convergence rate and reduce residual mean square error. The efficiency of the MCMA is proved via computer simulation of underwater acoustic channel equalization.