时变信道的自适应均衡算法比较

Comparison of adaptive equilibrium algorithms using time varied channel

针对时变信道需进行自适应均衡的要求 ,分析了常用的 3种基本自适应算法 :迫零算法 (ZF) ,最小均方算法(LMS)及递归最小二乘算法 (RLS) .迫零算法由于其原理上的缺点 ,实际应用很少 ;具有同样收敛速度和跟踪性能的递归最小二乘算法要优于最小均方算法 ,但前者运算量较大 ,程序复杂 ,且一些算法不稳定 .文中讨论的算法 ,虽然是针对线性、横向均衡器引入的 。

The paper analyzes three basic adaptive algorithms according to the requirement of time varied channel: ZF(Zero Force), LMS(Least Mean Square) and RLS(Recursive Least Square), and compares their characteristics and application respectively. ZF is rarely used because of its defect in principle. RLS algorithm that has the same convergence speed and track performance precedes LMS. But the former needs much calculation and the program is complicated. In addition, it may be unstable for some RLS. The algorithm is also applicable to other equalizer structure although they are introduced for liner and transversal equalizer.