Subspace approach to robust adaptive beamforming
Li Zou · Applied Acoustics · 2008
The data received by an array are severely distorted due to the uncertainty of the array itself and the nonstationary in the shallow water waveguide. The traditional adaptive beamformings such as the minimum-variance distortionless response (MVDR) are extremely sensitive to these mismatches.In this paper,a robust adaptive subspace beamforming is presented.It is assumed that the signal of interest belongs to a known linear subspace but that its coordinates within this subspace are otherwise unknown,and the interferences are also in some subspaces. First,the unknown parameters are estimated using maximum-likelihood estimator. Next,the maximum-likelihood estimates are used to derive a generalized likelihood ratio test (GLRT).The GLRT detector is called subspace beaforming.The performance of the subspace beaforming is illustrated by means of simulations under the conditions of several kinds of mismatches and by the experimental in-sea data. The results of simulations and experimental data show that subspace offers an improved performance on estimation and detection and is more robust than MVDR.