Chair: Jean-Jacques Fuchs, IRISA (FRANCE)
J.J. Smith, Curtin University of Technology (AUSTRALIA)
Y.H. Leung, Curtin University of Technology (AUSTRALIA)
A. Cantoni, Curtin University of Technology (AUSTRALIA)
The eigenvector method for estimating the positions of the receivers of an ocean- towed array is based on the eigendecomposition of the array signal correlation matrix to find the phase delays between the array receivers. Previous work has shown that for reasonable SNR, the bias and variance of the phase estimates is relatively independent of the number of receivers in the array. This suggests that the computational cost of the eigenvector method could be substantially improved by partitioning the array receivers into groups of smaller sub-arrays and then applying the eigenvector method to each sub-array. This paper introduces the interleaved partitioned eigenvector method, and shows that it significantly reduces the computational cost without adversely affecting the quality of the position estimates. Numerical examples substantiate the theoretical work of this paper and also demonstrate the improvement in beamforming when employing the shape estimation algorithm.
John R. Buck, Woods Hole Ocean Institute (USA)
James C. Preisig, Woods Hole Ocean Institute (USA)
Mark Johnson, Woods Hole Ocean Institute (USA)
Josko Catipovic, Woods Hole Ocean Institute (USA)
An algorithm is presented to excite a single mode in a shallow water channel using a vertical source array controlled by feedback from a reference hydrophone array. The algorithm iterates between computing the source weights based on its current estimate of the mode coupling in the channel, and updating its estimate of that coupling based on the modes observed at the feedback reference array. This allows us to excite high fidelity modes with confidence at a given location. The ability to control these modes depends on the accuracy with which they are observed. To this end, we compute the error for the linear least-squares mode estimator for scenarios where the feedback array does not span the entire water column. Finally, we present preliminary results obtained in a laboratory wave guide illustrating the successful convergence of the algorithm in a physical experiment.
Joseph Tabrikian, Naval Undersea Warfare Center
Hagit Messer, Tel-Aviv University (ISRAEL)
Source localization in a waveguide involves a multi-dimensional search procedure. In this paper we propose a new algorithm, in which the search in the depth direction is replaced by polynomial rooting. The proposed algorithm decreases the search dimension to one for a 2D localization problem (range and depth) and to two for a 3D one (range, depth, and direction-of arrival (DOA)), independently of the number of sources. Consequently, the presented algorithm requires significantly less computation.
Brian F. Harrison, Naval Undersea Warfare Center
Richard J. Vaccaro, University of Rhode Island (USA)
Donald W. Tufts, University of Rhode Island (USA)
Matched-field source localization methods attempt to estimate the range and depth of a source in an acoustic waveguide. These methods give good results when the waveguide parameters are known precisely; however, matched-field methods have been shown to be very sensitive to model mismatch resulting from errors in the assumed environmental parameters. In this paper, we describe an approach which minimizes model mismatch by estimating the environmental parameters and source location together. We propose an efficient way to initialize the maximum-likelihood search by projecting the received data onto subspaces corresponding to regions in parameter space.
V. Premus, Duke University (USA)
D. Alexandrou, Duke University (USA)
A technique is presented for the estimation of a set of parameters associated with a geologically motivated model for seafloor microroughness due to Goff and Jordan. The method seeks to connect the spatial covariance of the backscattered acoustic field with the correlation properties of the seafloor by constructing the a posteriori probability density function (pdf) of the parameters that define the seafloor microroughness wavenumber spectrum. The processor maximizes the joint a posteriori probability density of the model parameter set. Due to the complexity of the probability surface, the method of simulated annealing is used to search for the globally optimum solution vector.
C.F. Mecklenbrauker, Ruhr University Bochum
D. Maiwald, FGAN/FFM
J.F. Bohme, Ruhr University Bochum (GERMANY)
In this paper, we address the simultaneous detection and classification of signal arrivals, as well as the estimation of source parameters of signals impinging on an array of sensors. We develop a procedure for short-time stationary broadband signal propagation in a shallow ocean. The procedure will be applied to sensor data obtained from a towed horizontal receiver array in the Baltic Sea for interpreting the impinging signals. We apply a multiple test procedure to three propagation models: a Green's function model, uncorrelated normal modes and plane waves. We apply an F-test which calculates signal to noise ratios for the three models, and select the model with highest SNR. Thereby, we are able to classify previous phantom bearing estimations as uncorrelated normal mode propagation originating from the towing ship itself.
Robert N. Carpenter, Naval Undersea Warfare Center
Steven M. Kay, University of Rhode Island (USA)
A common problem in active sonar is that of multipath propagation in the ocean environment. A generalized likelihood ratio test (GLRT) approach is developed for detecting multiple target returns with unknown time delays and amplitudes in reverberation. The reverberation is modeled by a scattering function that is assumed to be described by the power spectral density of a complex autoregressive process. The generalized likelihood ratio test detector is described and its performance is compared to the clairvoyant optimal processor and an ad hoc processor.
O.P. Kenny, Defence Science and Technology Organization (AUSTRALIA)
L.B. White, Defence Science and Technology Organization (AUSTRALIA)
A robust detection statistic for signals whose parameters are uncertain is presented in this paper. Standard detection schemes generally use time domain correlation which can be related to correlation based on the Wigner-Ville distribution by Moyal's identity. This paper shows that a more robust detection statistic can be obtained by using generalised time-frequency space and deriving a non-linear time domain correlation. The performance of the robust detection statistic is evaluated with the aid of receiver operating curves. The paper gives two examples which shows that the robust detector under nominal conditions does not perform as well as standard detection schemes but gives better performance when the parameters of the signal deviates away from its nominal conditions.
Matthew A. Dzieciuch, University of California- San Diego (USA)
Ocean acoustic tomography has enjoyed success at shorter ranges (<2Mm) by using ray theory to model the acoustic propagation. As the ranges of these experiments increase to basin and global scale it is necessary to consider the dis- persive effect of the medium by using mode theory to model the propagation. This is useful at shorter ranges as well to correctly interpret the final cutoff which contains a substantial amount of energy for axial transmissions.
P. Blanc-Benon, Thomson-Sintra (FRANCE)
G. Bienvenu, Thomson-Sintra (FRANCE)
Target Motion Analysis (TMA) is a basic function in passive SONAR, generally using bearings only or bearings and frequency measurements. But due to the arrays whose apertures are practically negligible considering the target range, and even if the platform moves itself to yield a "synthetic array," the classical TMA methods take a few minutes to give an acceptable solution. Hence, this paper presents an enhanced TMA estimator using jointly the bearings and multipath parameters: the differential time-delays and their doppler shifts. The Cramer-Rao are studied for two cases of sound propagation: a constant celerity profile and a bilinear one. They both exhibit advantages in terms of a shorter time to get a given precision on the target parameters: its range, depth, and speed vector.