水声信道盲均衡的最小平方峭度恒模算法

Least Square Kurtosis Constant Modulus Algorithm for Blind Equalizer in Underwater Sound Channel

用误差信号峭度定义了平方峭度代价函数 ,提出了盲均衡器权系数更新的最小平方峭度恒模算法 ,该算法更新方程中含有的误差信号峰度因子有效地消除了高斯性误差信号的影响 ,加快了收敛 ,减小了收敛后的均方误差和码间干扰。用负声速梯度水声信道 ,对算法的性能进行了仿真研究。结果表明 :该算法在收敛速度 ,收敛后的均方误差及码间干扰等方面的性能优于常数模算法与最小平均峭度恒模算法。

To overcome the disadvantages of slow convergent rate and large residual mean square error (MSE) of constant modulus algorithm (CMA) and least mean kurtosis CMA, a square kurtosis cost function is defined as a kurtosis factor of error signals and its performance is analyzed, a least square kurtosis CMA for updating weight vectors of blind equalizer is proposed. The kurtosis factor based on error signals can improve convergence rate and make the proposed algorithm converge to global minimum. Thus the algorithm has a much faster convergence rate than that of the least mean kurtosis CMA and the CMA. The efficiency of the method is proved by computer simulation of negative acoustic gradient in underwater sound channel.