基于压缩感知的线性调频雷达成像

LFM Radar Imaging Based on Compressive Sensing

在经典的压缩感知理论中,最常用的测量矩阵是高斯随机矩阵和伯努利随机矩阵,它们具有最优的约束等容性质(R IP).与其他测量矩阵相比,精确重建同样的稀疏信号所需的采样数最少。然而完全的随机性导致它们在具体工程应用中受到限制,同时大大增加了重建算法的时间和空间复杂度。基于统计约束等容性质(StR IP)和确定性测量矩阵理论提出一种基于线性调频信号的压缩感知雷达成像方法,利用雷达发射的线性调频信号来构造测量矩阵,并用仿真实验证明了利用该测量矩阵可以很好地实现稀疏信号的重建。

In the classical compressive sensing theory,the most widely used sensing matrices are Gaussian and Bernoulli random matrices,which have optimal Restricted Isometry Property（RIP） and the number of samples needed for accurate reconstruction of sparse signals is minimal comparing with other types of sensing matrix.However,the randomness limited their use in practical engineering applications,and it also has a negative impact on the complexity,in both time and space,of reconstruction algorithms.A CS radar imaging method by transmitting Linear Frequency Modulated（LFM） signal is proposed based on the theory of StRIP and deterministic sensing matrices.The sensing matrix is constructed using the LFM signal,and the 1D simulation result show that this kind of deterministic sensing matrix works well for sparse signal reconstruction.