采用时频聚集性较好的线性调频信号作为线性时不变系统输入激励，采用Gabor字典作为过完备原子库。在利用传统系统辨识法之前先利用稀疏分解算法将输出信号进行去噪处理，显著提高系统辨识精度。 具体包括互谱算法，信号的Gabor稀疏分解的详细代码-Space can be a time for sparse decomposition to solve the problem of huge memory needed。This approach, combined with the rapid d

本文分别以稀疏基有离散余弦变换基（DCT）和快速傅立叶变换基（FFT）做为稀疏基，高斯随机矩阵、部分哈达玛矩阵为测量矩阵，L1范数、正交匹配追踪算法（OMP）为重建算法进行压缩感知算法实现。(In this paper, DCT and FFT are used as sparse basis, Gauss random matrix and partial Hadamard matrix are used as measurement matrix, L1 norm and OMP are u