Rayleigh信道下的支持向量机多用户检测方法  

A SUPPORT VECTOR MACHINE BASED DETECTION METHOD ON RAYLEIGH CHANNEL

作　　者：杨恒[1] 张贤达[1] 

机构地区：[1]清华大学自动化系智能技术及系统国家重点实验室,北京100084

出　　处：《电子与信息学报》2002年第2期257-260,共4页Journal of Electronics & Information Technology

基　　金：国家自然科学基金(69772023)

摘　　要：在BPSK调制的DS-CDMA中,基于支持向量机(Support Vector Machine,SVM)的多用户检测方法采用支持向量机的分类方法将接受向量分成+1和-1两类,达到检测的目的。与MMSE方法不同的是,支持向量机分类器的目的是找出一个能将训练向量中信号为+1和信号为-1的两类数据分离的最佳分离超平面。从数值仿真结果可以看出,在Rayleigh信道,这种支持向量机的多用户检测方法与MMSE多用户检测器相比,输出能达到较低的误码率。In DS-CDMA system with BPSK modulation, Support Vector Machine (SVM) based multiuser detection uses SVM classification method to classify received vectors into two classes. One is the received vectors of desired user's symbol +1 and the other is the vectors of desired user's symbol - 1. So desired user's symbol can be detected by this method. Different with MMSE method, SVM classifier finds the optimal separating hyper/plane that separates the class of +1 and class of - 1. Simulation results show that the performance of the SVM detector is better than that of MMSE detector in Rayleigh channels.

关 键 词：RAYLEIGH信道 支持向量机 多用户检测 码分多址系统 

分 类 号：TN914.53[电子电信—通信与信息系统]