Research On 2D DOA Estimation Methods Based On Planar Square Matrix

In the research of array signal processing,Direction of Arrival(DOA)as a key aspect has been researched extensively over years.The research of one-dimensional DOA occupies a larger proportion for its representativeness and luminous comprehension.But in fact,two-dimensional DOA estimation has more practical significance,such as providing accurate source location without blind areas for base stations and high-precision DOA tracking for moving targets to achieve better spatial multiplexing.Most of one-dimensional DOA estimation can be simply extended to two-dimensional ones for spatial location estimation.However,simple extension will undoubtedly bring faults,such as additional computation,low efficiency of solutions,and inapplicability of certain algorithms.Hence,in view of the above problems,this thesis carries out the following research:Firstly,this thesis combines the major challenges in implementing the two-dimensional DOA estimation based on URA-poor accuracy in low signal-to-noise ratio and snapshot deficiency environment,and analyzes the advantages and drawbacks of common narrow-band DOA estimation algorithm like MUSIC,ESPRIT and WSF in two-dimensional application.Simulation results show that the above DOA estimation algorithm have drawbacks in aspects of threshold performance,with WSF of best accuracy but lowest computational speed.Secondly,aiming at the problem of poor threshold performance,pseudo-noise resampling technology is introduced into WSF algorithm.In order to solve the consequent problem of large computation and slow solution,fireworks algorithm is combined and improved in the process of WSF solution,and a PR-WSF algorithm based on the improved fireworks algorithm is proposed.Simulation results show that the improved algorithm enhances the threshold performance and possesses higher speed.However,the requirement for large amounts of snapshots becomes the key element in hindering the further improvement of solution speed.Thirdly,given that the DOA estimation algorithm based on compressed sensing theory has low sensitivity to signal-to-noise ratio and requires less snapshots in accurate estimation of DOA information to obtain high speed in processing less data,this thesis researches the compressed sensing DOA,and studies the two-dimensional direction of arrival estimationmodel with separable array manifolds,by which the two-dimensional off-grid algorithm based on first-order Taylor expansion is presented to reduce the coupling between the elevation dimension and the azimuth dimension in the array manifold matrix.Simulation results show that this method can effectively estimate the off-grid deviation.Finally,combined with SVD decomposition and multitask model,a 2D multitask compressed sensing off-grid algorithm based on SVD decomposition is presented,which improves the drawbacks of current algorithms in the problems of the great number of snapshots and enhance the speed effectively.Simulation results show that this algorithm is effective in solving both on-grid and off-grid problems,and possesses a better estimation performance in condition of low signal-to-noise ratio and snapshot deficiency.