DOA Estimation Under Impulsive Noise Based On Bounded Nonlinear Function And Sparse Representation

As a hot topic in array signal processing,direction of arrival(DOA)estimation plays an important role in many fields and applications,including wireless communication,vehicle localization,radar system and medical signal processing.Up to now,most researches focus on solving array data under Gaussian noise.However,with the development of data communication,the electromagnetic environment is becoming more complex.Influenced by manual interference and natural factors,the received array data may be affected by impulsive noise,like communication system,underwater acoustic system,sudden burst and sharp spikes.Conventional methods show performance degradation under impulsive noise and existing methods for impulsive noise still have some problems.Considering this problem,this dissertation pay attention to solve DOA estimation under impulsive noise,including non-circular signals' DOA estimation,sparse representation based DOA estimation and MIMO radar DOA estimation.The main research contents and contributions of this dissertation are as follows:(1)To solve non-circular signals' DOA estimation under impulsive noise,an extended array output is constructed by non-circular signal's structure and a bounded nonlinear covariance is built and proved to have a similar format to conventional covariance matrix.Furthermore,based on the proposed covariance matrix and space spectrum,we propose some improved methods to realize far-field and near-field DOA estimation under impulsive noise.Meanwhile,to better evaluate the performance for near-field DOA estimation under impulsive noise,the Cramer-Rao bound is derived.Otherwise,considering impulsive noise and co-channel interference,a generalized cyclic correlation is proposed based on bounded nonlinear function.Combing the proposed correlation and Toeplitz reconstruction,coherent and non-coherent signals DOA estimation is solved under impulsive noise.Based on these works,the proposed methods can significantly improve the performance for far-field and near-field DOA estimation under impulsive noise.(2)To solve sparse representation based ULA DOA estimation under impulsive noise,a series of methods is proposed.Inspired by the suppression on impulsive noise of Cauchy loss function,a Cauchy function based sparse representation is proposed to realize DOA estimation under impulsive noise and mutual coupling.Considering mixed-field DOA estimation under impulsive noise,a novel method is proposed utilizing correntropy based sparse representation and bounded nonlinear function.Considering wideband DOA estimation under impulsive noise,an improved method is proposed based on sparse representation and cyclic correntropy,hyperbolic tangent based cyclic correlation.Furthermore,to realize DOA estimation with unknown signals'number,a real-valued sparse Bayesian learning is proposed which shows better performance and lower computation.(3)To solve MIMO radar DOA estimation under impulsive noise,the monostatic MIMO radar is applied as signal model.For on-grid condition,a logarithmic hyperbolic cosine function based cost function is proposed in sparse representation,and the iteration steps are give.Based on this model,the performance for DOA estimation is improved.For off-grid condition,a sparse Bayesian learning based method is proposed to deal with MIMO DOA estimation,which shows better performance.