一种用于MIMO系统动态信道估计技术研究.pdfVIP

22 4 Vol22 N o 4 2008 8 JOURN AL OF ELE CTRON IC M EA SUREM EN T AN D IN STR UM EN T 39 M IMO 王 燚 郭 伟 (, 610054) : M IMO , M IMO ( SVR) M IMO ( AMSVR), MSVR, , M IMO : M IMO, , , : TN9 11 : A : 5105015 Research on Dynam ic ChannelE stim ation in M IMO Comm unication System W ang Y i Gu W ei ( Nati nalKey Lab fUESTC, Chengdu, 610054, Ch ina) Abstract: Channel estmi ati n is essen tial in d igital c mmun icati ns. But n nfrequency and stati nary fading channel is the prerequ isite f rM IMO transm issi n n w adays, wh ich lmi its the usage fM IMO technique main ly t ind r c mmunicati n. In th is paper an adaptive mu ltip le d mi ensi nal supp rt vect r regress r ( AM SVR) is pr p sed t estmi ate the M IMO c mmun icati n channels based n the g d generalizati n pr perties f sh rt train ing sequences. Instead f using quadratic pr gramm ing (QP) pr cedures f r s lving theM SVR, this paper presents an iterative rew eighted least squares pr cedure t mi pr ve the train ing c nvergence speed fM SVR, wh ich is a pr cedure smi ple t mi p lem ent and all w s a recursive f rmu lati n. The perf rm ance f the pr p sed alg rithm is veri fied using c mputer smi u lati ns in M IMO c mmun icati n scenari s. K eywords: M IMO system, supp rt vect r m ach ine, channel estmi ati n, mu ltiregressi n , , 1 , M I MO, M IMO M IMO ( BLAST ) , M IMO, , , [ 1]