非确定性MIMO系统的K-Best检测算法研究

Research on K-Best Algorithm for Rank-Deficient MIMO System

研究无线信道优化问题,由于非确定性MIMO系统的信道矩阵行不满秩特性,导致MMSE-GDFE左预处理后引入了强干扰,使得传统K-Best算法产生了误码平层效应。为了消除误码平层效应,提出部分最大似然和K-Best结合算法,并在初始检测的强干扰区域利用部分最大似然算法获得局部最优解,降低强干扰对系统的影响消除误码平层效应,进而在局部最优解的基础上利用K-Best算法完成剩余符号的检测,保证MIMO检测器的软输出特性。仿真结果表明算法消除了传统K-Best算法中的误码平层效应,保证了检测的可靠性。

The characteristic of non-full row rank of channel matrix for rank-deficient MIMO systems introduces severe interferences which result in error floor effect in traditional K-Best Algorithm with left preprocessing based on MMSEGDFE.With the motivation to remove the unfavorable effect,the combination of partial ML and K-Best Algorithm is proposed.The partial ML Algorithm is used in the initial strong-interference region to find the partial optimal solution,and then the traditional K-Best Algorithm is exploited to detect the rest symbols based on this solution to keep the soft-output character of the MIMO detector.The simulation results show that the error floor effect can be eliminated with the proposed method,the different tradeoff between performance and computational complexity can be reached by adjusting the number of symbols in partial ML Algorithm and the optimal number is L=Nt-Nr＋1.