基于改进的PASTd子空间跟踪的半盲多用户检测-Journalof.PDFVIP

第2 8 卷第5期 东 北 大 学 学 报 ( 自 然 科 学 版 ) V ol28, No. 5 2 00 7 年 5 月 Journal of Nort heastern U n vers ty( N atural Sc ence) M ay 2 0 0 7 PASTd 孟 艳, 汪晋宽, 宋 昕, 韩英华 ( , 110004) : , , PAST d , , PAST d , , , , PA ST d OPAST , SV D , : CDM A ; ; ; PAST d ; M M SE : T N 911. 7 : A : 1005- 3026( 2007) 05- 0648- 04 Sem-i blind Mult-i user Detection Based on Improved PASTd Subspace Tracking Algorithm M EN G Yan , WA N G J in-kuan, S ON G X in , HA N Ying-h ua ( School of Infor mat on Sc ence Eng neer ng , N ortheastern U n vers ty, Shenyang 110004 , Ch na. Correspondent : M ENG Yan , E- ma l : ymeng @ ma l . neuq. edu . cn) Abstract: Several algor thms of subspace t rack ng are nvest gat ed. T he e genvalue decompos t on (EV D) and s ngular value decompos t on ( SV D) are not su ted for eng neer ng mplement at on because of t he r h g h comput at on com plex t es, w h le t he PAST d algor thm br ngs about very slow convergence rate because of t he non-ort hogonal t y of est m at ed e genvectors though t s comput at on complex t y s low . A n mproved PA ST d subspace t rack ng algor thm s theref ore proposed t o be appl ed to the subspace- based sem-bl nd mult-u ser det ector for adapt ve subspace est mat on. T he algor thm proposed can nsure t he orthogonal ty of t he e genvectors, thus qu cken ng ts convergence rate. S mulat on results show t hat t he proposed algor t hm s super or to bot h the PA ST d sem- bl nd mult- user det ect on and OPAST sem- bl nd mult-u ser detect on n convergence rate, b t error rat e ( BER) and outpu