An Research To Robust Adaptive Beamforming Algorithms

The Capon Beamformer,as one of well-known adaptive beamformer,is to maximize the output signal-to-interference-plus-noise ratio(SINR)and simultaneously maintain a distotionlss response toward the desired signal.Theoretically,it assumes that there is no desired signal component in the beamforming training data and the steering vector of desired signal is known precisely.However,such assumptions are far from realistic since there exists a lot of errors in practical applications,and the beamformer will suffer severe performance degradation.Addressing this problem,this thesis reaserches and proposes several novel robust adaptive beamforming algorithms to reject these existing errors.The main work and innovation of the thesis covers these points as following:1.Describe the background and research significance of robust adaptive beamforming algorithms,and review the development history and research status at domestic and abroad.2.Researched the array data model and fundamentals of adaptive beamformer,and analyze the effect caused by array element position error and covariance matrix error for the MVDR beamformer in theory.It also introduces several classic robust adaptive beamforming algorithms and discusses their advantages and disadvantages in this thesis.3.Addressing the mutual coupling problem,a robust adaptive beamforming algorithm based on mutal coupling coefficients estimation is proposed.The proposed algorithm only needs the array geometry,and can jointly estimate the Derection-ofArrivals(DOA)and mutual coupling coefficients,which is a closed-form solution.The simulation result verified the proposed algorithm can achieve high output SINR and array gain in the presence of mutual coupling4.Addressing the desired signal in training data and steering vector mismatch problem,the thesis proposed an improved robust beamforming algorithms based on the previous research.It reconstructs the interference-plus-noise covariance matrix based a sptial power spectrum sampling method and estimates the desired signal steering vector by adding a new quadratic constraint,then combine them to calculate the weight vector.The proposed method avoids the pow dense function estimation,only needs a little prior information and has low complexity.Through the simulation,it shows that the proposed algorithm has superior performance than the tested beamforming algorithms in different error scenarios such as DOA error,array perturbation,coherent local scattering,incoherent local scattering and gain-phase error.