Research On Robust Hybrid Beamforming For Large Scale Arrays

Adaptive beamforming is an important spatial filtering algorithm in array signal processing.In practical scenarios,when faced with various error factors such as steering vector errors and covariance matrix errors,beamforming performance will be severely degraded.Therefore,research on robust adaptive beamforming is necessary.And with the increasing array size,a hybrid digital analog array has been proposed to solve the cost problem of large-scale arrays.In previous studies,there was little focus on the robust beamforming problem of mixed digital analog arrays under the MVDR criterion.Committed to the research of robust adaptive beamforming for large-scale mixed digital analog arrays,the main work and innovative points of this article are as follows:1)In order to combat the small sample problem caused by large-scale arrays,a small sample robust adaptive beamforming algorithm is proposed in the case of narrowband signals.This algorithm utilizes a standard Gaussian random matrix to construct a projection matrix,which is used for dimensionality reduction of sample covariance matrix and guidance vector.The dimensionality reduction sample covariance matrix and guidance vector are used to obtain more accurate Capon power spectrum,and to reconstruct the interference plus noise covariance matrix(INCM).The steering vector of the desired signal is estimated by solving the Quadratic programming(QCQP)problem.Finally,the adaptive weight vector is calculated based on the reconstructed INCM and the estimated guidance vector.This algorithm has better performance compared to other algorithms when the number of samples is less than the number of array elements.2)In the broadband signal scenario,the stability of Minimum Variance Guided Adaptive Beamforming(STMV)and MVDR(also known as ISM algorithm)at each frequency point was compared through statistical analysis.A robust STMV algorithm was designed to address the steering covariance matrix(STCM)error and steering vector error in the STMV algorithm.This algorithm mainly estimates a more accurate expected signal steering vector for each frequency point through the robust beamforming(RCB)criterion,and then constructs an accurate pre steering matrix for each frequency point.Then,using the STMV broadband spectral estimation method,the expected signal power was estimated to reconstruct the interference plus noise oriented covariance matrix.This algorithm can effectively combat wideband signal steering vector error.3)A specific solution method is proposed for the non convex problem of optimizing the beamforming weights of mixed digital analog arrays under the MVDR criterion.Firstly,the optimal all digital beamforming weight vector is obtained using the previously proposed robust algorithm.For the fully connected structure,the compressed sensing framework is introduced to sparsely represent the analog and digital weights,and the orthogonal Matching pursuit algorithm is used to solve the sparse basis to obtain the analog weights; In a partially connected structure,the structure of the simulated beamforming matrix in the partially connected structure is utilized to obtain the optimal solution through alternating iterative optimization.In the case of broadband,still use pre directed processing and solve using the above method.The proposed method can approximate the performance of fully digital beamforming.