Robust adaptive matched-field localization based on a subspace projection distance estimator

The matched field localization algorithm in the presence of uncertainties in the ocean environment based on the replica field noise subspaces perturbation constraints is presented. In each searching grid, the environmental parameters are randomly sampled and the replica field vectors are computed, the replica field covariance matrix is formed with these replica field vectors and the eigenvalue decomposition (EVD) is performed. Using eigenvectors with relatively small eigenvalues, the constraint matrix is obtained. The same process is performed for covariance matrix of the measured data, and the eigenvector with the largest eigenvalue is used as the signal vector. The localization ambiguity surfaces are obtained with the constraint matrix and the signal vector. With defining a probability of correct localization (PCL) and Peak-to-Background Ratios(PBR), the performance of the suggested algorithm is researched for different environmental perturbation and constraint matrix dimension using the simulation data, which are derived from MFP workshop held in 1993 at the Naval Research Laboratory(NRL), and the experimental data, which are derived from the Mediterranean Sea. The Results show that the suggested algorithm is robust.