Abstract: A parallel interference cancellation QR decomposition detection algorithm of VBLAST-OFDM is proposed.It makes hypothesis to the first detection layer and divides several parallel branches.The detected signal is eliminated successively from the receive vector of each branch，after that，other signals are detected by using QR factorization，but only the last detected signal that has the highest diversity gain is reserved.Finally，the optimal solution of several branches is selected to be as the final detection result by using partial ML and the performance is improved effectively.Simulation results show that the proposed P-ICQR algorithm is superior to the conventional parallel algorithm，IQRD algorithm，QR algorithm and ZF algorithm.

Key words: Vertically Bell Laboratories Layered Space-Time（VBLAST）, Miltiple Input and Multiple Output（MIMO）, QR algorithm, P-ICQR algorithm

摘要： 提出了一种VBLAST-OFDM系统中的平行干扰消除QR分解检测算法，称为P-ICQR算法。该算法首先对最先检测层信号做出假设，分成多个平行分支，在每个分支上依次干扰消除已检测信号的影响，运用QR分解判决余下层信号而只保留分集增益最高的最后检测层判决信号，最后用部分最大似然法对平行分支选取最优解作为最终检测结果，有效提高了系统的检测性能。仿真结果表明，提出的P-ICQR算法比传统的平行算法、循环迭代QR分解算法、QR算法、迫零算法的误码性能都要好。

关键词: 分层空时码, 多输入多输出（MIMO）系统, QR算法, P-ICQR算法