基于支持向量拟合的恒模盲均衡算法

Blind Equalization of Constant Modulus Based on Support Vector Regression

本文提出了一种利用迭代最小二乘支持向量拟合(IRWLS-SVR)的盲均衡算法。该算法利用数字通信系统中调制方式的恒包络特性(CMA),在基带条件下对具有严重码间干扰的接收信号进行均衡处理,实现无线信道的无误差传输。该方法不同于以往基于神经网络算法和高阶统计量的盲均衡方法,具有收敛速度快,干扰抑制强,计算量小的优点。计算机仿真结果表明,经过多径瑞利信道衰减后的接收信号可以通过该均衡器得到正确解调。

In this paper a CMA algorithm for blind equalization based on IRWLS - SVM is proposed. The algorithm uses the characteristics of common modulated amplitude in digital communication system, equalizes the received baseband symbols with serious distortion from the ISI aroused in wireless channel with muhipath and Doppler frequency shift, and realizes error-free communication in wireless channel. Compared with other algorithms based on neural network or higher order cummulations, it has the advantages of rapid convergence, robust to interference and small calculation burden. Simulation results show that the received signal attenuated by muhipath Rayleigh channel can he demodulated using the proposed equalization algorithm correctly.