Research On Sparse Robust Super-resolution DOA Method

The array error in reality is unavoidable.Whether it is the traditional DOA estimat io n algorithm or the compressed sensing method,most of the DOA estimation algorithms will be seriously affected by the array error,which will greatly reduce the estimation effect even invalidate the algorithms.To solve this problem,this paper studies the robust DOA estimation algorithm in the presence of array errors.The main research work of the paper are:1,The array error model in DOA estimation is introduced.The influence of two kinds of errors,the amplitude and phase errors and the error of the array element's position,on the array manifold pattern is mainly analyzed.The influence of array element position error can be projected on the phase error.This provides a theoretical model basis for the next study of a robust DOA estimation method.2,Then for the problem that the array error has a great influence on the DOA estimat io n results,this paper first studies the rank-one correlated model to suppress the influence of noise;then in the context of compressed sensing and sparse frame representation,based on this rank-one correlated model,using null space optimized sparse recovery techniques,studies a robust DOA estimation method,null space tuning algorithm,and analyzes the complexity and precision of the algorithm.Finally,a large number of simulation experiments are carried out to verify the effectiveness of the method.The success rate and RMSE of the estimation results under different array types,different signal-to-noise and snapshots are simulated,and the simulation experiments and performance under different signal types are simulated.The method has both robustness,high resolution,noise robustness,and valid it y of the estimated number of signal sources.3,In order to improve the freedom of the array,the paper studies and analyzes two kinds of sparse arrays: the coprime array and the nested array.The related concept of the difference Co-array are studied,and two kinds of coprime arrays,basic and extended coprime arrays,are studied.And the degree of freedom that can be gained by their respective difference Co-array are analyzed in detail.The two level nested array are analyzed.The K level nested array model are optimized,using find a method of how to design the nested array to maximize its degree of freedom when the number of sensors is given;in order to fully utilize the advantages of sparse array and the null space tuning algorithm: the null space tuning algorithm based on coprime array and nested array is proposed,.And a large number of simulation experiments are performed respectively.First,the availability of the algorithm is verified,when the number of sources is more than the number of sensors.Then,the success rate and RMSE of the estimation results under different SNR,snapshot number,amplit ude error and phase error are simulated.A large number of simulation experiments show that the null space tuning algorithm based on the coprime array and the nested array is a robust super resolution DOA estimation algorithm.