基于最小错误概率变步长的盲均衡算法

Variable Step Size Blind Equalization Algorithm Based on Minimum Error Probability

在分析了基于最小错误概率盲均衡算法的基础上,用牛顿梯度变步长算法实现了基于最小错误概率新的变步长盲均衡,克服了固定步长收敛速度和收敛精度之间的矛盾,加快了收敛速度,减小了稳态剩余误差。计算机仿真结果分别给出了4PAM,8PAM信号在典型电话信道和普通最小相位信道中牛顿梯度变步长算法与固定步长算法的收敛曲线。由两种算法收敛曲线的比较可以看出,新算法的收敛性能明显得到改善。

This paper analyzes a blind equalization algorithm based on minimum error probability.The new minimum error probability blind equalization based on Newton gradient steepest ascent variable step size algorithm is proposed.Computer simulations demonstrate the proposed algorithm accomplished faster convergence and steady smaller MSE than fixed step size algorithm.