Evaluation of an orthogonal beamforming procedure using real data

We report on the application of an orthogonal beam-forming procedure to real data received at a linear array of equally spaced hydrophones in shallow water. This is a high resolution method to determine the directions of incoming wavefronts. With simulated data it has often been shown to yield improved angular resolution compared to conventional techniques. Here we present results concerning the resolution capability, number of detectable targets, and processing time when it is applied to real passive SONAR data. We report on the limited applicability and reliability of multivariate statistical tests which are proposed to achieve the proper separation of the coherent signal contributions from the total signal, which is the method's crucial step. We conclude that these tests have to be modified in order to cope with real data.