Research On High-resolution Array Signal Processing Technique

Array signal processing has been an important branch of modern communications, and it is widely used in radar and sonar. It plays an irreplaceable role on target detection and localization. Due to the urgent needs of national defense construction, we need high-resolution array signal processing technology to achieve precise localization about the target in the sea and in the air. Moreover, the real-time acquisition and reliability of radar and sonar must be guaranteed for military requirement, which requires that the high-resolution signal processing technology must have good robustness and timeliness.Various robust algorithms are improving the robustness of high-resolution array signal processing. With the advent of the theory of signal reconstruction, the theory of Compressive Sensing provides a new solution for timeliness. This article mainly study high-resolution array signal processing technology based on linear array, and the theory of Compressive Sensing and robust algorithm are applied to the array signal processing.Firstly, the article does the research on multiple signal classification algorithm and its improved algorithms, and does performance analysis about the factors that affect the performance. Because coherent signals are difficult to distinguish, the article does the research on multiple signal classification algorithm for decorrelation based on matrix rebuild and analyzes its performance. Besides, the paper does the research on multiple signal classification algorithm based on the pre-processing data by array element to improve the robustness of the conventional multiple signal classification algorithm and overcome the shortcoming of the high SNR threshold. Meanwhile, the method can enhance the input signal noiseratio and azimuth resolution. The multiple signal classification algorithm for coherent and noncoherent signals is proposed for broadband signal and analyzes its performance.Secondly, another classic high-resolution algorithm, ESPRIT, is studied. the paper researches the standard ESPRIT and its improved algorithms. Then the article does performance analysis about the factors that affect the performance and compare the performances of these algorithms, which lets us know the good azimuth estimation algorithm.Furthermore, the article takes two robust algorithms to improve the robustness of the algorithm based on the adaptive algorithm of the sample covariance matrix inversion. The two robust algorithms are the diagonal loading algorithm and weighted vector norm constraint algorithm. And the paper tells us the good algorithm by comparing the robustness performance of these algorithms.Finally, the article makes use of the spatial sparsity of array signal to use compressive sensing technique for the DOA estimation modeling, And the space is divided by uniform angle and uniform sine. The classical greedy algorithm, the orthogonal matching pursuit algorithm, is used to solve the problem of compressed sensing. The performance analysis to a single snapshot data and multiple snapshots data is made.