一种基于交叉验证的稳健SL0目标参数提取算法

Cross validation based robust-SL0 algorithm for target parameter extraction

利用雷达目标在空间的稀疏特性,研究了一种基于压缩感知的伪随机频率步进雷达(compressive sensing based pseudo-random step frequency radar,CS-PRSFR)。首先,在分析CS-PRSFR目标回波的基础上,建立了目标参数提取模型;然后,针对在噪声统计特性未知时,传统稀疏信号重构算法无法适用的问题,提出一种基于交叉验证的稳健SL0(robust SL0based on cross validation,CV-RSL0)目标参数提取算法。CS-PRSFR由于其感知矩阵较强的非相关性,可获得更高的距离-速度联合分辨性能;该算法无需已知噪声统计特性,随着信噪比的提高,其目标参数提取性能能够快速逼近最佳估计的下限。仿真结果表明该方法的正确性和有效性。

Utilizing the space sparsity property of radar targets,a compressive sensing based pseudo-random step frequency radar（CS-PRSFR） is studied.Firstly,the CS-PRSFR targets echo is analyzed and the targets parameter extracting model is constructed.To solve the problem of inapplicability of traditional sparse signal reconstruction algorithms amid noise of unknown statistics,a cross validation based robust SL0（CV-RSL0） algorithm extracting the parameter of targets is proposed.Because of the better incoherence of the sensing matrix,the CS-PRSFR can obtain a higher range-velocity joint resolution performance.The proposed algorithm needs no prior information of the noise statistics,and the performance of its targets parameter extraction can rapidly approach the lower bound of the best estimator as the signal to noise ratio improving.Simulation results illuminate the correctness and efficiency of this method.