DOA Estimation Based On Compressed Sensing In Non-Gaussian Noise Environment

Direction of Arrival (DOA) estimation is an important content in array signal processing, its main task pass an array of multi-sensor data, and estimated to the spatial distribution of the direction of arrival of the signal source.In recent years, with the development of Compressed Sensing theory, people take advantage of Compressed Sensing theory sparse to estimate the DOA based on sparse reconstruction method,and it can solve the shortage of traditional DOA estimation method.Currently,DOA estimation method based on Compressed Sensing sparse reconstruction mostly studied under Gaussian noise background,and the nature of some of the noise often has a strong pulse property,and can not be described by a Gaussian distribution.In Non-Gaussian noise conditions, the performance of DOA estimation method declined based on Gaussian noise background.Therefore,the study of the DOA estimation method is of great significance based on the Compressed Sensing sparse reconstruction theory in the Non-Gauss noise environment.Alpha stable distribution has excellent properties,it can be well described Non-Gaussian pulse noise.The fractional lower order statistics as an important tool to analyze the Alpha stable distribution.In this paper, Alpha stable distribution as mathematical model,combine the fractional lower order statistics theory to Compressed Sensing theory.Study on Compressed Sensing DOA estimation method based on fractional lower order statistics.Firstly,using vector of array received signal fractional lower order correlation matrix and vector of array received signal phase fractional lower order covariance matrix has a spatial sparsity characteristics.Based on the fractional lower order correlation matrix vector sparse reconstruction DOA estimation algorithm(FLOM-VEC) and phase fractional lower order correlation matrix vector sparse reconstruction DOA estimation algorithm(PFLOM-VEC) are proposed.In order to better suppress the influence of impulse noise on reconstruction Algorithm,by analyzed vector signal characteristics,Improved fractional lower order correlation matrix vector sparse reconstruction DOA estimation algorithm (IFLOM-VEC) and Improved phase fractional lower order correlation matrix vector sparse reconstruction DOA estimation algorithm(IPFLOM-VEC) are proposed.As for these four algorithms we referred to as the Fractional lower order statistics vector sparse reconstruction DOA estimation algorithm(FLOS-VEC).Finally,simulation of these algorithms and array covariance matrix vector sparse reconstruction DOA estimation algorithm (C-VEC).The simulation results show that the performance of FLOS-VEC algorithm is much better than the C-VEC algorithm in the Alpha stable distribution noise environment. In other words,FLOS-VEC algorithm can better suppress the Non-Gaussian pulse noise.