Research On Underwater DOA Estimation Method Based On Compressed Sensing

As an important research Direction of array signal processing,Direction of Arrival(DOA)estimation technology has an important application and development in underwater target positioning and other fields.With the continuous development of Marine resources,underwater DOA estimation plays an increasingly important role.Marine noise has the characteristics of color noise in the spectrum distribution,and has the signal coherence and different power spectrum characteristics compared with the general white noise.As the Gaussian white noise model is generally used for underwater DOA estimation,there will be a large error in the actual estimation.On the other hand,the traditional DOA estimation technology needs to meet the requirements of the minimum sampling frequency for signal acquisition,which brings a certain burden and resource waste to the storage space and transmission process of the sampled signal.Existing compressed sensing theory provides a solution to the problem of signal sampling,which can ensure that the sparse signal can be fully recovered after sampling at a lower frequency.However,it fails to effectively solve the problem of underwater environmental factors affecting the estimation performance.In this paper,combined with the theory of sparse reconstruction compressed sensing and underwater DOA estimation algorithm,an in-depth study on the influence of white noise and color noise in the underwater environment will be conducted.The main work of this paper is as follows:1.The theoretical basis of underwater DOA estimation and the theory of sparse reconstructed compressed sensing are introduced,including the mathematical model of signals,common arrays of one-dimensional and two-dimensional DOA estimation,DOA estimation model and reconstruction principle of sparse reconstructed compressed sensing,as well as the measurement index of DOA estimation performance.2.The estimation principle and specific steps of the classical sparse reconstruction compressed sensing algorithm are analyzed,and the one-dimensional and two-dimensional CRB estimation based on uniform linear array are derived respectively.The estimation performance and operational complexity of the classical algorithm are simulated and analyzed.The simulation results show that the L1-SVD algorithm has better estimation accuracy in the estimation performance.3.In view of the impact of low signal-to-noise ratio and colored noise on the estimation performance of the algorithm,a one-dimensional DOA estimation algorithm(IUW-L1-SVD)with iteratively updated weights is proposed.The weighted matrix of the sparse optimal solution was constructed based on the weighted idea,and the coefficients of the weighted matrix were iteratively updated.In order to reduce the impact of noise on the estimation accuracy,the dimension reduction matrix was de-noised before convex optimization.The simulation results show that the proposed algorithm has higher probability of correct estimation and lower estimation error in low SNR white noise and color noise environment,and the algorithm complexity is moderate.4.In view of two-dimensional DOA estimation,a two-dimensional DOA estimation algorithm(2D-IUW-L1-SVD)with iteratively updated weights is proposed.In this algorithm,a double dictionary observation matrix is constructed instead of a single dictionary observation matrix to carry out convex optimization,and Angle matching is carried out based on the spatial spectral amplitude matching method,so as to realize the estimation of two-dimensional Angle.The simulation results show that the algorithm has better estimation performance and good suppressing effect on noise.