| Array direction finding technology,also known as direction of arrival(DOA)estimation,is widely used in military and civilian fields such as radar positioning,earthquake monitoring,mobile communication and so on.However,with the increasingly complex electromagnetic environment and the increasing demand for direction finding technology,the traditional DOA estimation technology has many defects and is difficult to meet the established requirements.In recent years,combining compressed sensing(CS)theory with DOA estimation technology and using sparse reconstruction algorithm to determine the direction of incoming wave signal,it can better adapt to the harsh environment such as low signal-to-noise ratio and less snapshots.It also has good advantages in dealing with coherent sources,reducing sampling rate and low-cost signal transmission processing.Therefore,this paper mainly studies the related contents of DOA estimation under the framework of compressed sensing theory:Firstly,the sparse DOA estimation models under equal angle and equal sine spatial meshing are studied.Because the virtual array flow pattern matrix constructed by grid serves as the measurement matrix in the sparse DOA estimation model to realize the compressed sampling of the signal,the orthogonality performance analysis is carried out for the measurement matrix constructed by the above two different spatial grids,which paves the way for the subsequent selection of measurement matrix;DOA estimation is usually analyzed and solved under the multi observation model(MMV).In order to reduce the computational complexity under the MMV model,the array covariance matrix is vectorized through Khatri Rao product,and a DOA estimation model based on the sparse representation of covariance vectorization is constructed.Finally,the MMV problem is transformed into a single observation vector(SMV)problem.Secondly,the smoothing L0 norm reconstruction algorithm is studied,and a weighted composite optimization smoothing L0 norm algorithm(WCOSL0)is proposed.By introducing the composite optimization function with higher approximation L0 norm degree,the accuracy of the algorithm is improved,and the weight function is introduced to accelerate the acquisition of sparse solution and improve the convergence of the algorithm.Then,in order to adapt to the low SNR environment,the regularization mechanism is introduced to improve the projection direction in the inner loop of the proposed WCOSL0 algorithm.Finally,a robust weighted composite optimization smoothing norm reconstruction algorithm(RWCOSL0)is proposed and used in the constructed DOA sparse estimation model to convert the MMV model into SMV model.Simulation results show that the proposed algorithm has better DOA estimation performance in non ideal environments such as low signal-to-noise ratio and less snapshots. |
