Towed array shape estimation using Kalman filters-experimental investigations

The problem of estimating the shape of a towed array instrumented with either depth sensors, compasses, or both in a discrete-time state-space formulation is treated in a companion paper by D. A. Gray et al. (to appear), in which the state-space representation is derived from a dynamical model of the propagation of tow-point-induced motion down the array. A Kalman filter is derived to recursively estimate the shape of this towed array, and solutions to the Riccati equation are used to predict the mean square error of the Kalman filter array shape estimates. The present study investigates the performance of this Kalman filter approach as an array shape estimator using both simulated examples and sea trial data. Fundamental to the Kalman filter approach is the model that describes the dynamical behavior of the towed array. The results of an experimental program that was undertaken to validate this model are also presented