Performance evaluation of horizontal and vertical vector sensor arrays in shallow water environments

Abstract

In planning an experiment it is essential to have as diverse information as possible in order to obtain a good estimate of the relevant parameters. Traditionally, the inversion for waveguide parameters as well as the source location is done exclusively using pressure measurements. Recently, it has been suggested by Nehorai and Paldi [IEEE Trans. Signal Process. 42, 2481–2491 (1994)] that arrays of vector sensors (sensors that measure both pressure and velocities) could be used. Experiments have also been conducted in deep water by D’Spain et al. [IEEE Oceans (1992)]. In a realistic shallow-water environment the use of either vector or pressure sensors on a linear horizontal/vertical array with amplitude-only or complex-valued data at a number of discrete frequencies are discussed. Based on developed maximum-likelihood functions, the performance for the above type of data will be classified using either a global or a deterministic approach. For the deterministic approach, the Cramer Rao lower bound (CRLB), based on derivatives of the likelihood function, is used. For the global approach the a posteriori distributions are examined. These are based on Monte Carlo integration of the likelihood function. [Work supported by DSO National Laboratories, Singapore.]