多天线系统中逼近最大似然的快速检测算法

An Efficient Detection Algorithm for MIMO Systems Approaching ML Performance

该文针对无编码的多输入多输出无线通信系统中的最大似然检测接收机在发端天线数较多、调制阶数较高时计算复杂度过高的问题,提出了一种低复杂度的球形译码算法。该算法首先利用信道信息对待检测的发送信号矢量进行分组,然后对各组内的信号矢量采用球形译码进行最大似然检测,并在组间做干扰消除。理论分析和仿真表明,该算法不仅复杂度低,而且能够逼近最大似然检测的性能。

An efficient Sphere Decoding （SD） algorithm is proposed for uncoded Multi-Input Multi-Output （MIMO） systems with a large number of antennas and higher modulation constellations. This method utilizes a channel-based group detection strategy. The transmitted symbols on each substream are detected by performing sphere decoding within each group and successive interference canceling between groups. Analyses and simulations demonstrate that the proposed algorithm can approach Maximum Likelihood （ML） detection with a relatively low complexity.