摘要：
In the MIMO(multiple-input multiple-output) communication system, QRD-M (QR decomposition associated with the M-algorithm to MLD) algorithm has a good application prospect because it employs QR decomposition and constellation points screening of each layer to accomplish low complexity signal detection.However, when the number of the transmitting and receiving antennas increases, it is difficult to achieve a good tradeoff between system performance and implementation complexity.In order to solve the problem, this paper proposes a novel signal detection method by using SNR(signal-noise ratio) sorting.Based on the breadth-first parallel structure of QRD-M algorithm, the proposed method first sorts the values of SNR of receiving antennas to avoid the error propagation and therefore speeds up the tree search process.Combining with dynamic threshold search method, the search radium is reduced until reaching the branch of the minimum accumulated squared Euclidean distances.The simulation results show that the proposed method can approach the performance of the maximum likelihood detection with lower complexity.%在多输入多输出(multiple-input multiple-output,MIMO)系统信号检测中,基于虚实分解的宽度优先检测算法(QR decomposition associated with the M-algorithm to MLD,QRD-M)通过QR分解和对每层星座点的筛选,实现了较低复杂度的检测,具有很好的应用前景.但该算法随收发天线数和调制阶数的增加而难以实现性能与复杂度的折衷.针对此缺点,提出了一种基于信噪比排序的信号检测改进方法.该方法在传统QRD-M算法的基础上,通过对不同接收天线进行信噪比(signal-noise ratio,SNR)排序,从信噪比最大的天线开始检测,避免了误差传播现象,从而加速树搜索过程,再结合动态门限树搜索,不断缩小搜索半径,直至找到最小累计度量值所在分支.仿真结果表明,与传统QRD-MLD算法相比,基于性噪比排序的动态门限信号检测算法能以较低的复杂度获得接近于最大似然检测的性能.
