基于一种新的S型函数快速凸组合最小均方算法

为解决传统凸组合自适应滤波算法在联合参数迭代计算量大、算法收敛速度慢、跟踪性能差等问题,提出了一种基于一种新的S型函数快速凸组合最小均方(SCLMS)算法;该算法用一种新的S型函数,代替Sigmoid函数,在保证和CLMS算法相同稳态误差情况下,避免了指数运算,减少了计算量;同时也提高了收敛速度和信号的跟踪性能。通过独立高斯白噪声作为输入信号算法仿真、相关噪声作为输入信号算法仿真;以及非平稳环境下算法仿真;并对三种仿真结果进行了分析,验证了该算法性能可靠有效。

A fast convex combination least mean square algorithm based on a new S type function

In order to solve the calculate in large quantity, slow convergence speed and tracking performance is poor in the joint parameter iteration in traditional convex combination of adaptive filtering algorithm, this paper proposes a new fast convex combination of LMS algorithm based on type S function (SCLMS).Using a new S function of the algorithm, instead of Sigmoid in function, when the SCLMS steady-state error is same with CLMS algorithm , avoid the index operation, reduce the amount of computation, and improve the convergence speed and tracking performance of the signal. The independent Gauss white noise as input signal simulation algorithm, noise as input signal simulation algorithm, and non-stationary environment algorithm simulation, and three kinds of simulation results were analyzed and verified the performance of the algorithm is reliable and effective.