一种新的并行检测算法下的信道分组方法

A New Method of Channel Partition on Parallel Detection

摘要:

在发射天线数为 NT、接收天线数为 NR的 MIMO系统中,并行检测算法可达到最优的检测性能,但现有系统中用到的信道分组方法是计算 NT次矩阵广义逆来完成信道分组,随着天线数量的不断增加,矩阵求逆计算量会快速增加,从而进一步加剧并行检测算法计算复杂度。针对该问题,文章证明了:在并行检测算法中,只须一次矩阵广义逆的计算即可完成信道分组。仿真结果表明,应用新的信道分组方法的格归约并行检测算法的性能不变。

Abstract:

In MIMO system with NT transmitting antennas and NR receiving antennas, the parallel detection algorithmscan achieve optimal detection performance. However, the channel partition method used in the prior art is to calculate the NT times of the pseudo-inverse of the channel matrix to complete the channel partition. With the increasing number of an?tennas, the complexity of the pseudo-inverse of the channel matrix will increase rapidly, which further aggravates the com?putational complexity of parallel detection algorithms. For this problem, in this paper it’s proved that in parallel detectionalgorithms, only one calculation of the pseudo-inverse of the channel matrix was needed. To complete channel partition.Simulation results show that the performance of the lattice reduction parallel detection algorithm applying the new channelpartition method is unchanged.