Robust Adaptive Beamforming Under Data Dependent Constraints

As an important branch of array signal processing,the robust adaptive beamforming(RAB) has been widely used in many fields.In the traditional robust adaptive beamforming algorithm,a common problem is that the estimated guidance vector of the desired signal will converge to the interference subspace,so that the desired signal is suppressed as interference,and the interference is outputted as the desired signal without distortion.Those problem above leading the degradation.The following research works are carried out in this paper to improve the robustness of the adaptive beamformers to various error factors:First of all,when the Capon power spectrum was used to reconstruct the signal covariance matrix,a power threshold value was introduced to eliminate the signal components that could not be SOI,so as to reduce the error of the reconstructed signal covariance matrix.Secondly,we use the oblique projection method to reconstruct the interference covariance matrix.This method can eliminate all the SOI components in the received signal,and ensure that the interference signal will not be affected almost,so that make the accuracy of the reconstructed interference covariance matrix is higher.At the same time,the sampling in the angle range of interference signals is avoided,the computational complexity of reconstruction is reduced,and the covariance matrix is reconstructed in a easier way.Subsequently,a Fractional Quadratically Constrained Quadratic Program(FQCQP) optimization problem is proposed based on the reconstructed signal covariance matrix and interference covariance matrix.It is found that the optimization problem we proposed is a typical non-convex quadratic constrained quadratic programming problem(QCQP),which can be transformed into a positive semidefinite optimization problem(SDP),and then solved by using the CVX toolbox.By taking advantage of the fact that the rank of the signal covariance matrix is relatively small and the scale and phase change of the steering vector has no influence on the optimization problem and constraint conditions,the optimization problem initially proposed by us is simplified,and finally our problem is transformed into an optimization problem with linear constraints.We use Lagrange multiplier method combined with Newton's method to solve the optimization problem we proposed.Besides,the time complexity of the optimization problem can be greatly reduced by using the method we mentioned before.Finally,the above two methods are used to solve the optimization problem proposed in this paper.It is found that the solution results are exactly the same by comparing the simulation results.The method proposed in this paper is lower than the CVX method in terms of time complexity.In addition,we compare the robust adaptive beamforming algorithm proposed in this paper with some other classical robust adaptive beamforming algorithms by simulation.The results show that our algorithm is outperforming the previous algorithms in most aspects.