摘要
提出了一种新的采用结构风险最小化(SRM)准则的盲均衡器(SRM-BE).该方法根据信号的特征恢复思想,以SRM为准则构造具有时间去相关特性的代价函数,采用静态迭代学习算法在线跟踪信道.通过仿真实验,并与采用最小均方误差准则的盲均衡器(LMS-BE)和采用神经网络的盲均衡器(NN-BE)进行比较,结果表明该方法的非线性均衡性能最佳.
A new blind equalizer based on structure risk minimum （SRM） rule is presented. According to the feature reconstruction of signals, the method constructs the time decorrelation cost function under SRM rule, and traces channel using static iterative learning algorithm. Simulation is carried out to compare it with least mean square-based blind equalizer （LMS-BE） and neural networks-based blind equalizer （NN-BE）, and the results show the method has much better performance.
出处
《北京邮电大学学报》
EI
CAS
CSCD
北大核心
2008年第4期6-9,共4页
Journal of Beijing University of Posts and Telecommunications
基金
中国博士后科学基金项目(20070421094)
国家自然科学基金项目(60772056)
关键词
均衡器
盲均衡
结构风险最小化
特征恢复
equalizer
blind equalization
structure risk minimum
feature reconstruction
