No abstract

Matched-field based methods always involve the comparison of the output of a physical model and the actual data. The method of comparison and the nature of the data varies according to the problem at hand, but the result becomes always largely conditioned by the accurateness of the physical model and the amount of data available. The usage of broadband methods has become a widely used approach to increase the amount of data and to stabilize the estimation process. Due to the difficulties to accurately predict the phase of the acoustic field the problem whether the information should be coherently or incoherently combined across frequency has been an open debate in the last years. This paper provides a data consistent model for the observed signal, formed by a deterministic channel structure multiplied by a perturbation random factor plus noise. The cross-frequency channel structure and the decorrelation of the perturbation random factor are shown to be the main causes of processor performance degradation. Different Bartlett processors, such as the incoherent processor ͓Baggeroer et al., J. Acoust. Soc. Am. 80, 571-587 ͑1988͔͒, the coherent normalized processor ͓Z.-H. Michalopoulou, IEEE J. Ocean Eng. 21, 384 -392 ͑1996͔͒ and the matched-phase processor ͓Orris et al., J. Acoust. Soc. Am. 107, 2563-2375 ͑2000͔͒, are reviewed and compared to the proposed cross-frequency incoherent processor. It is analytically shown that the proposed processor has the same performance as the matched-phase processor at the maximum of the ambiguity surface, without the need for estimating the phase terms and thus having an extremely low computational cost.

No abstract

This paper considers the inversion of experimental field data collected with light receiving systems designed to meet operational requirements. Such operational requirements include system deployment in free drifting configurations and a limited number of acoustic receivers. A well-known consequence of a reduced spatial coverage is a poor sampling of the vertical structure of the acoustic field, leading to a severe ill-conditioning of the inverse problem and data to model cost function with a massive sidelobe structure having many local extrema. This causes difficulties to meta-heuristic global search methods, such as genetic algorithms, to converge to the true model parameters. In order to cope with this difficulty, broadband high-resolution processors are proposed for their ability to significantly attenuate sidelobes, as a contribution for improving convergence. A comparative study on simulated data shows that high-resolution methods did not outperform the conventional Bartlett processor for pinpointing the true environmental parameter when using exhaustive search. However, when a meta-heuristic technique is applied for exploring a large multidimensional search space, high-resolution methods clearly improved convergence, therefore reducing the inherent uncertainty on the final estimate. These findings are supported by the results obtained on experimental field data obtained during the Maritime Rapid Environmental Assessment 2003 sea trial.

No abstract

One of the most stringent impairments in matched-field processing is the impact of missing or erroneous environmental information on the final source location estimate. This problem is known in the literature as model mismatch and is strongly frequency dependent. Another unavoidable factor that contributes to model mismatch is the natural time and spatial variability of the ocean waveguide. As a consequence, most of the experimental results obtained to date focus on short source-receiver ranges (usually <5 km), stationary sources, reduced time windows and frequencies generally below 600 Hz. This paper shows that MFP source localization can be made robust to time-space environmental mismatch if the parameters responsible for the mismatch are clearly identified, properly modeled and (time-)adaptively estimated by a focalization procedure prior to MFP source localization. The data acquired during the ADVENT'99 sea trial at 2, 5, and 10 km source-receiver ranges and in two frequency bands, below and above 600 Hz, provided an excellent opportunity to test the proposed techniques. The results indicate that an adequate parametrization of the waveguide is effective up to 10 km range in both frequency bands achieving a precise localization during the whole recording of the 5 km track, and most of the 10 km track. It is shown that the increasing MFP dependence on erroneous environmental information in the higher frequency and at longer ranges can only be accounted for by including a time dependent modeling of the water column sound speed profile.

Abstract-Matched-Field Processing (MFP) is now a mature technique for source localization and tracking. There are at least two aspects that emerge, by their relevance, to the success of MFP: one is the ability of a given MFP processor to accurately pinpoint the source location while rejecting sidelobes, and the other is the impact of erroneous or missing environmental information (known as model mismatch) in the final source location estimate.This study addresses the first aspect regarding sidelobe rejection while considering that the processor is working on a mismatch free situation. One well known procedure to reduce sidelobes is to use a broadband MFP processor (whenever a band of frequencies is available). There are a number of different ways to combine MFP information across frequency that can be classified in two broad groups: the conventional incoherent methods, that are based on the direct averaging of the auto-frequency inner products and the, say, less conventional methods, that perform a weighted average of the cross-frequency inner products where the weights are the frequency compensation phase-shifts. The later are generally termed as coherent broadband methods since they combine complex inner products.The coherent broadband methods proposed in the literature are either suboptimal or very computationally intensive, even for a small number of frequencies. An alternative method is presented that combines cross-frequency information with the same localization performance than the standard coherent methods and a computation load similar to that of the incoherent processor. The performance of the various broadband processors is compared in simulated data.

No abstract