混沌支持向量机优化小波加权多模盲均衡算法  

Chaos and Support Vector Machines Optimization Based Wavelet Transform Weighted Multi-modulus Blind Equalization Algorithm

作　　者：郭业才[1] 徐文才[1] 许芳[1] 

机构地区：[1]南京信息工程大学电子与信息工程学院,南京210044

出　　处：《系统仿真学报》2013年第3期451-454,459,共5页Journal of System Simulation

基　　金：全国优秀博士学位论文作者专项资金(200753);安徽省高等学校自然科学基金(KJ201 0A096);江苏省自然科学基金(BK2009410);2012年度高校科研成果产业化推进项目(JHB2012-9);江苏省六大人才高峰基金(2008026);江苏省高校优势学科"传感网与现代气象装备"

摘　　要：为了提高均衡器对高阶QAM信号的盲均衡性能,提出了基于混沌支持向量机优化的小波加权多模盲均衡算法。为避免权向量陷入局部极小值点,该算法利用支持向量机对均衡器的权向量进行初始化,并用混沌优化算法对支持向量机参数进行优化;利用正交小波变换对均衡器的输入信号进行预处理,来降低输入信号的自相关性;采用加权多模算法来调整权向量迭代过程中的模值。水声信道的仿真结果表明,与加权多模盲均衡算法和小波加权多模盲均衡算法相比,该算法具有较快的收敛速度和更小的稳态误差。In order to improve the blind equalization performance for high-order QAM signals, chaos and support vector machines optimization based wavelet transform weighted multiple modulus blind equalization algorithm was proposed. To avoid local minima, the proposed algorithm utilized support vector machines to initialize the weight vectors of equalizer. During the process, the parameters of support vector machines were optimized by chaos optimization algorithm. The equalizer＇s input signals were pretreated by using orthogonal wavelet transform to reduce the self-correlation. The proposed algorithm adapted the Weighted Multiple Modulus Algorithm （WMMA） to adjust the modulus of weight vectors. The simulation results with the underwater acoustic channel, compared with WMMA and WT-WMMA （Wavelet Transform based WMMA）, show that the proposed algorithm has higher convergence speed and smaller residual mean square error.

关 键 词：盲均衡 正交小波变换 支持向量机 混沌优化 加权多模算法 

分 类 号：TN911.7[电子电信—通信与信息系统]