STABILIZED HIGH-RESOLUTION BEAMFORMING WITH HORIZONTAL ARRAYS - 2 EXPERIMENTAL TRIALS IN SHALLOW-WATER

High-resolution algorithms are potentially able to outperform conventional linear methods at the task of beamforming, i.e., spatial-spectrum estimation, and in controlled experiments or simple computer simulations they often do prove superior. In marine environments, however, high-resolution techniques frequently fall short of these expectations. Their outputs can take on an unstable appearance characterized by fading of the true target peak and by the occurrence of multiple false targets. In recent years a class of stabilization techniques has evolved to moderate these effects by suppressing contributions from the smaller eigenvalues of the cross-spectral density matrix. While such techniques generally produce good results in simulations with the maximum likelihood estimator and even with the more sensitive maximum entropy estimator, they have seen little use on sea data. Here, results are presented of two sea trials of a simple eigenvalue stabilization procedure in a relatively demanding class of environments - shallow water. The technique provided what is considered to be marginal stabilization for the maximum likelihood estimator and none at all for maximum entropy.