Research On Self-correction Technology For Gain-phase Error In Polarization-sensitive Array

In the field of array signal processing DOA estimation,compared to traditional scalar arrays that can only receive the spatial information of the incident signal,the polarization-sensitive array can obtain both the spatial and polarization domain information simultaneously,with more robust detection capabilities,stronger interference suppression ability,and higher system resolution.At present,most researches on signal processing of polarization-sensitive arrays mainly focus on multi-parameter estimation,and do not consider array errors.In practical engineering applications,polarized arrays are affected by various errors(gain-phase error,sensor location errors,mutual coupling and so on),and the performance of most polarization high-resolution algorithms will be seriously degraded or even invalid.Therefore,it is necessary to compensate or correct the array errors.Based on the polarization-sensitive uniform circular array,this paper studies t he self-correction technology for gain-phase errors.The main research contents are as follows:(1)Extend classical iterative self-correction algorithm from scalar array to polarization domain,and propose a DOA-polarization-error reduced-dimension iterative self-correction algorithm.First,DOA-polarization and error are separated,and each iteration is divided into estimated error parameters and estimated DOA-polarization joint parameters;Then,a reduced-dimension MUSIC algorithm based on matrix rank loss is used to jointly estimate the DOA and polarization parameters in each iteration,the four-dimensional search is optimized into a two-dimensional search only related to the azimuth and elevation angle,and directly calculate the polarization parameters using the search results;Finally,when the DOA-polarization parameters are fixed,the problem is transformed into a quadratic extremum problem,and the estimation of the gain-phase errors is completed.The algorithm can converge after multiple iterations.(2)The multi-dimensional parameter joint self-correction algorithm based on the maximum likelihood criterion is studied.The cost function constructed by the criterion is essentially a multi-dimensional nonlinear optimization problem.This paper uses two solutions: the first is to use the gradient-based Gauss-Newton method combined with the rotation projection algorithm to solve,and proposes a GN-ML self-correction algorithm.The second is the intelligent optimization algorithm.This paper innovatively uses the covariance matrix adaptive evolution strategy to optimize and solve,and proposes the CMAES-ML self-correction algorithm.The simulation results show that the the algorithm has achieved good parameter estimation performance.(3)The multi-dimensional parameter joint self-correction algorithm based on the principle of weighted signal subspace fitting is studied.This principle is similar to the maximum likelihood criterion in the field of DOA estimation.The constructed cost function is also a multi-dimensional nonlinear optimization problem,which is solved by Gauss-Newton method and covariance matrix adaptation evolution strategy,The GN-WSF self-correction algorithm and the CMAES-WSF self-correction algorithm are proposed,and good parameter estimation performance is achieved.the efficiency of each self-correction algorithm is verified through computer simulation experiments.