The Signal Parameter Estimation Based On Compressive Sensing

The compressive sensing(CS) based signal parameter estimation could break the limitation of this theorem, it only needs a small mount data of data to complete the high-resolution or super-resolution estimation of signal parameters for sparsity signals. The thesis makes a detailed introduction of the CS theory form three aspects, its mathematical model, compression and measurement, signal reconstruction. It mainly focuses on the sparsity parameter estimation of sparse signal and the DOA estimation of array signal to introduce the application of compressive sensing signal in parameter estimation. The mainly work and innovations are listed as follows:1. This thesis proposed a new signal reconstruction method called AStMP which could estimate the sparsity of the sparsity signals accurately based on the research of the existing sparsity reconstruction algorithms. It described the AStMP in improvements of the initial estimation of the sparsity, the adaptive step-size, atom pre-selection and the support set, and verifies the advantages of AStMP through MATLAB simulations.2.The thesis studies three kinds of CS-based DOA estimation algorithms: 1l-SVD,OMP and FOCUSS. It demonstrates their advantages to the traditional DOA estimation algorithms in the aspects of the DOA estimation peak and the influence of snapshot, signal-to-noise ratio, the number of receiving array parameters to the accuracy of the calculate direction of arrival values by MATLAB simulation. Based on the analysis of 1l-SVD and OMP, combines the thought of SVD and OMP, and does the QR decomposition of the dictionary vector during the calculation of the sparse solution and the updating of the residual in OMP to reduce the computational complexity. The simulation results prove that such algorithm performs better than the original OMP in DOA estimation, could inhabit the influence of noise while lower the computational complexity.