基于正交匹配追踪的压缩感知信号检测算法

Compressive sensing signal detection algorithm based on orthogonal matching pursuit

在不重构信号的情况下,利用检测算法直接处理压缩感知信号获得的采样值可以完成信号检测任务。目前的检测算法中,作为判决依据的特征量仅由一次迭代过程决定。在感兴趣信号存在时,算法获得的特征量波动较大,影响检测性能。基于这种原因,本文提出一种改进算法,在每次迭代中基于正交匹配追踪思想修正特征量,降低特征量的波动,提高检测性能。实验结果表明,与原算法相比,本文算法获得的特征量波动小,相同检测阈值下,检测成功率高,相同检测成功率要求下,需要的采样点数少,允许的信噪比低。

Without reconstructing the signal itself,signal detection could be solved by a detection algorithm,which directly processes the sampling values obtained from compressive sensing signal.In current detection algorithm,as the judgment criterion,the characteristic quantity is determined in one iterative process.When the signal of interest exists,the characteristic quantity obtained from the detection algorithm fluctuates,which decreases the detection performance.Therefore,in this paper,we propose an improved algorithm,which amends the characteristic quantity based on orthogonal matching pursuit in each iteration,reduces the fluctuation of the characteristic quantity and improves the detection performance.Experiment results show that compared with original algorithm the proposed method greatly reduces the fluctuation of the characteristic quantity,has higher success rate of detection under the same detection threshold,requires fewer samples under the same success rate of detection and allows lower SNR under the same success rate of detection.