Study On The Method Of Robust Adaptive Beamforming In Antennas Array

Beamforming is also called as the spatial filtering,which is an important branch of array signal processing and has been extensively applied in radar,sonar,seismology,wireless communications and other areas.When there is no error,traditional adaptive beamforming algorithms can precisely receive the desired signal and suppress the interference and noise signals.However,there are array geometry,propagation environment,signal direction and other errors in real systems,which will lead to the performance degradation in traditional adaptive beamforming algorithms or even the signal cancellat ion phenomenon.Therefore,the robustness to errors for adaptive beamforming algorithms has to be improved.The main contents of this dissertation are listed as following:Eigenspace-based beamforming algorithms need to assume that the dimension of the subspace is exactly known.Under low signal-to-noise ratio(SNR)conditions,the performance of this class of algorithms will decrease.This dissertation firstly proposes a robust adaptive beamforming algorithm based on the reconstruction of the signal plus interference covariance matrix(SICM).This algorithm uses the properties of the vectorized array covariance matrix to construct the signal plus interference subspace,and projects the enhanced sample covariance matrix onto the constructed signal plus inter ference subspace to reconstruct the SICM.By constructing an accurate signal subspace,the desired signal steering vector can be estimated.This algorithm has robustness to the covariance matrix errors and steering vector errors.A robust adaptive beamforming algorithm using the cyclostationarity of the source signals is also proposed.This algorithm uses the cyclostationarity of the source signals to estimate the steering vectors of all signals and construct the SICM.This algorithm can avoid the step of constructing subspace and can avoid the problem of imprecise subspace dimension in constructing subspace.This algorithm can simultaneously improve the performance of eigenspace-based beamformers under the conditions of low and high SNRs.The adaptive beamforming algorithms based on interference covariance matrix reconstruction remove the desired signal component from the sample covariance matrix,which can overcome the desired signal cancellation phenomenon.However,the resolution and the accuracy of the Capon spatial power spectrum used in these algorithms are relatively low.This dissertation firstly proposes a low-complexity algorithm based on the iterative adaptive approach spectrum estimation.The iterative adaptive approach spatial spectrum with higher accuracy,higher resolution and more robustness to coherent signals and limited number of snapshots is used to replace the traditional Capon spatial spectrum.This algorithm can reduce the computational complexity of iterative adaptive approach algorithm and can obtain good beamforming performance.Then a knowledge-aided robust adaptive beamforming algorithm based on the iterative adaptive approach spectrum estimation is proposed.This algorithm can estimate the actual interference steering vectors and its corresponding powers.Therefore,this algorithm can reconstruct precise interference plus noise covariance matrix(INCM)and output superior beamforming performance.Finally,an angular-sector robust adaptive beamforming based on iterative adaptive approach spectrum estimation is proposed.This algorithm determines the respective narrow angular sectors of the desired signal and the interference signals,and samples these narrower angular sectors.This algorithm uses an iterative adaptive approach algorith m to estimate the power corresponding to each sampling point.This algorithm can achieve good balance bwteen reducing the computational complexity and obtaining satisfactory beamforming performance.Uniform linear arrays are limited by the Nyquist sampling theorem.When the number of incident signals exceeds the number of elements in the antennas array,uniform linear arrays cannot provide enough degrees-of-freedom(DOFs)to suppress all interference signals.This dissertation firstly proposes a coprime array adaptive beamforming algorithm based on the interference covariance matrix reconstruction.By deriving the coprime virtual coarray,this algorithm can extremely increase the number of DOFs in antennas array.This algorithm can estimate the directions and powers of all incident signals by modifying the coprime virtual uniform linear array Capon spectrum estimator.The actual steering vector of each incident signal can be estimated by solving the mismatch vector of steering vector for each incident signal.Therefore,this algorithm can reconstruct precise coprime array INCM.Then a coprime virtual uniform linear iterative adaptive approach algorithm for robust adaptive beamforming is proposed.This algorithm accurately estimates the powers of the incident signals.Different from the existing coprime array beamforming algorithms,this algorithm can directly process the single-snapshot received signal vector of the virtual uniform linear array without repairing the rank of the virtual uniform linear array covariance matrix.Finally,a coprime virtual non-uniform iterative adaptive approach algorithm for robust adaptive beamforming is proposed.This algorithm uses coprime virtual non-uniform iterative adaptive approach algorithm to estimate the power spectrum of signals.Existing coprime array beamforming algorithms need to construct full-rank Toeplitz matrix and neglect the information in non-uniform virtual antennas.This algorithm can directly process the single-snapshot received data of virtual non-uniform array,and can fully utilize the information contained in all virtual array elements.When the directions of the incident signals are very close,the powers of the incident signals and the noise variance can still be accurately estimated in an iterative manner.Therefore,this algorithm can greatly improve the coprime array beamforming performance.Adaptive beamforming algorithms usually assume that the incident signal s are second-order circular signals.However,in the fields of radio communication and satellite communication,the incident signals are often non-stationary and second-order non-circular signals.In this dissertation,the non-circularity of non-circular signals is applied to the adaptive beamforming algorithms.A robust widely linear beamforming algorithm based on extended covariance matrix reconstruction and extended steering vector estimation is proposed,which can estimate the steering vectors,powers and non-circularity coefficients of all non-circular signals.Compared with existing widely linear beamforming algorithms,this algorithm can reconstruct more accurate INCM,pseudo INCM and extended INCM according to their definitions.Two extended subspaces containing the extended desired signal steering vector can be constructed.The extended desired signal steering vector can be estimated by obtaining the intersection set of these two extended subspaces.This algorithm can ensure that widely linear adaptive beamformer has satisfactory output performance and strong robustness under the conditions of various model mismatches.