基于稀疏表示的RAMP算法估计超宽带LFM信号参数

The Estimation of the Ultra-Wideband LFM Signals by Using the RAMP Algorithm Based on the Sparse Representation

针对奈奎斯特采样频率过高，硬件实现困难的问题，提出了基于压缩感知理论的超宽带线性调频信号的初始频率和调制斜率的估计方法。文章根据LFM信号的稀疏特征建立原子库，分别运用凸优化和RAMP算法对信号进行重构，实现对初始频率和调制斜率的估计。在仿真实验中对凸优化和RAMP两种重构算法的恢复效果进行比较，结果表明，凸优化算法重构准确度更高，但耗时较长；RAMP算法的信号重构准确度没有凸优化算法高，但耗时短，效率高。

In some applications, sampling with Nyquist frequency may be hard to implement due to hardware limitation. Based on compressive sampling（CS） theory, a method which estimates the initial frequency and modulation rate of the uhra-wideband LFM signal is presented in this paper. Based on the sparse characteristic of the LFM signal, the atomic library is established in this paper. The signal is reconstructed using the convex optimization and RAMP（ Regularized Adaptive Matching Pursuit） to realize the estimation of the initial frequency and modulation rate. In the simulation experiment, the results of two kinds of reconstruction algorithm are compared. The reconstruction algorithm of the convex optimization can get a more correct result than the reconstruction algorithm of RAMP. But it takes much more time to use the convex opti- mization. The reconstruction algorithm of RAMP works more efficiently.