ICA最优核窗估计的多用户检测  

Multi-user Detection Based on ICA with Processing of Optimal Bandwidth Kernel Estimatio

作　　者：宛明高[1] 李晓辉[2] 

机构地区：[1]池州学院机电工程系,安徽池州247000 [2]安徽大学电子科学与技术学院,安徽合肥230039

出　　处：《合肥师范学院学报》2012年第3期47-49,56,共4页Journal of Hefei Normal University

基　　金：安徽省高校省级自然科学研究项目No.KJ2012Z274;安徽省年度重点项目No.11070203010;池州市科技计划项目No.ZC09003~~

摘　　要：将核估计中递归选择窗宽的ICA算法应用于DS-CDMA系统的多用户检测中,在样本总体真实密度未知时,根据给定的核密度和抽取的样本点来选择最优窗宽。采用ICA最优核窗估计算法检测器的输出初始化独立分量分析的迭代,对任意混叠信号进行盲分离。通过仿真证明了该算法在DS-CDMA多用户检测中的有效性。ICA algorithm based on a bandwidth selection with recursive method for kernel estimation is applied to multi-user detection in DS-CDMA. When the true density in the sample population is unknown, the optimal bandwidth is selected according to the given kernel density and sample points. Any mixture signals are blindly separated by the initialization ICA iteration based the kernel window detector output. The simulation results show that the effectiveness of this algorithm in DS-CDMA multi-user detection.

关 键 词：最优窗宽 核密度估计 独立分量分析 多用户检测 

分 类 号：TN911.23[电子电信—通信与信息系统]