中文摘要：

大规模多输入多输出（multiple input multiple output,MIMO）中基站天线数远大于用户数,固匹配滤波（matching-filtering,MF）检测子在上行多用户检测中有重要的应用潜力。但是和传统的迫零（zero-forcing,ZF）、最小均方误差（minimum mean square error,MMSE）检测子相比,MF存在明显的性能损失,且在高信噪比区域出现＂误差门限＂效应。为了解决MF上述的不足,提出一种基于匹配滤波连续干扰相消和局部搜索的多用户检测算法。理论分析表明,与MF类似,所提方法的计算复杂度远低于MMSE、ZF,并且实验证明此方法的误符号率和误比特率明显低于MF。

英文摘要：

In large-scale multiple input multiple output（MIMO）,the antenna number in the base station reaches tens or even hundreds,which is much large than the user numbers.Consequently,matching-filtering（MF）detector can be applied for large-scale MIMO uplink multi-user detection（MUD）.However,compared to conventional zero-forcing（ZF）and minimum mean square error（MMSE）detectors,there is significant performance loss for MF.At the same time,＂error floor＂phenomenon is observed in the high signal-noise rate regions.In order to resolve these problems of MF referred above,a new MUD algorithm is designed based on matching filtering successive interference cancellation（MF-SIC）and local search.Theoretical analyses show that,similar as MF,the computational complexity of the new algorithm is one order lower than that of ZF and MMSE.Simulation results show that the new algorithm outperforms MF on both symbol error rate and bit error rate.