基于严格∪性价值函数的DF结构的盲均衡方案

A New Scheme of Decision-Feedback Blind Equalization Based on Strictly Convex Cost Function

本文提出了一种Ｄｅｃｉｓｉｏｎ－Ｆｅｅｄｂａｃｋ．（ＤＦ）结构形式的盲均衡价值函数．证明了该价值函数对前向均衡器系数矢量Ｃ，反向均衡器系数矢量Ｃ和系统传输误差概率Ｐｎ，都是严格的∪函数．并且证明了该价值函数保证收敛到全局极小点．分析和计算机模拟结果显示，对于典型１６ＱＡＭ形式的输入信号，本文提出的方案具有低的计算复杂度和快的收敛速度。

In this paper,a new decision feedback blind equalization cost function is introduced,which is strictly a convex function with respect to tap weight vector of forward equalizer,vector of the feedback equalizer and the error probability of system. Its global convergence is guaranteed by using gradient algorithm.For a typical channel input with 16QAM sequence,our scheme has proved to have low comoputational complexity and fast convergence rate.