广义RASK系统的回溯StOMP检测算法

本文提出了一种基于压缩感知（compressive sensing，CS）理论的低复杂度的检测算法:回溯分段正交匹配追踪（backtracking stagewise orthogonal matching pursuit，BStOMP）次优检测算法，较好地解决了广义接收天线移位键控（generalized receive antenna shift keying，GRASK）系统中最优最大似然（maximum likelihood，ML）检测算法会随着激活接收天线数的增多而计算复杂度大幅度增长的问题。该算法首先依据当前迭代残差和等效信道矩阵的内积选择出大于设定阈值的项作为候选激活天线索引；然后通过最小二乘法对候选激活天线索引进行初步估计；最后引入回溯过程进行筛选，剔除一些多余的索引，提高重构精度。仿真结果表明，在相同条件下，所提算法与传统StOMP检测算法相比，能在增加少量计算复杂度的情况下检测性能得到较大的提升，因此保证了在检测性能与计算复杂度间获得良好的平衡。 

Backtracking StOMP Detection Algorithm for GRASK Systems

? ?In this paper， a low-complexity detection algorithm based on compressed sensing （CS） theory is proposed: backtracking stagewise Orthogonal Matching Pursuit （BStOMP） suboptimal detection algorithm， which obviously solves the problem that the complexity of optimal maximum likelihood （ML） detection algorithm increases greatly with the increase of activated receive antennas in generalized receive antenna shift keying （GRASK） system. Firstly， when the inner product of iterative residuals and equivalent channel matrices is greater than the set threshold， the candidate active antenna index is selected. Then the candidate active antenna index is estimated by the least square method. Finally， the backtracking process is introduced to eliminate some redundant indexes and improve the reconfiguration precision. Simulation results indicates that under the same conditions， compared with traditional StOMP detection algorithm， the detection performance of the proposed algorithm can be greatly improved with a small increase in computational complexity， thus ensuring a good balance between detection performance and computational complexity.