The DOA Estimation Based On Compressed Sensing

Compressed sensing theory is proposed from the start has been of concern to the people, the theory overcomes the traditional Nyquist sampling theorem, using a small amount of sampling information for sparse signal or signals can be compressed exactly recovered, it has successfully solved the problem that the traditional method of large amount of data brought. With compressed sensing theory research unceasingly thorough, based on the application of compressed sensing of DOA estimation is also gradually mature, including the mesh complete signal sparse representation, some scholars was also made the sparse owe sampling of compressed sensing DOA estimation, using a small part of the array elements can be used to effect of full matrix when.In this paper, the problem is based on the compressed sensing DOA estimation, using the array sparse to complete the DOA estimates, in order to achieve or close to the effect of the time. Firstly, combined with sparse owe sampling model, were compared several different array sampling the pros and cons of methods, which have a good performance of the method, and combining wi th the method proposed some improvements on the sampling matrix, in order to get a better estimate of the effect.Then in order to get better estimation effect, using Bayesian compressive sensing theory, its application to the model, and the results and we often use compression sensing reconstruction algorithm OMP algorithm, analysis the method of estimation performance. Then, based on Bias compressed sensing, an improved algorithm is introduced. The improved algorithm of the Bias compressed sensing DOA is improved by SVD decomposition, and the better estimation results are obtained.. At last, a modified model of grid partition mismatch is established, and two different reconstruction algorithms are given, and the results of the two algorithms are compared with the real values.At last, this paper summarizes the problems of the next research.