Multipath arrivals at a receiving sensor are frequently encountered in many signal-processing areas, including sonar, radar, and communication problems. In underwater acoustics, numerous approaches to source localization, geoacoustic inversion, and tomography rely on accurate multipath arrival extraction. A novel method for estimation of time delays and amplitudes of arrivals with maximum a posteriori (MAP) estimation is presented here. MAP estimation is optimal if appropriate statistical models are selected for the data; implementation, requiring maximization of a multidimensional function, is computationally demanding. Gibbs sampling is proposed as an efficient means for estimating necessary posterior probability distributions, bypassing analytical calculations. The Gibbs sampler includes as unknowns time delays, amplitudes, noise variance, and number of arrivals. Through Monte Carlo simulations, the method is shown to have a performance very close to that of analytical MAP estimation. The method is also shown to be superior to expectation-maximization, which is often applied to time-delay estimation. The Gibbs sampling approach is demonstrated to be more informative than other time-delay estimation methods, providing complete posterior distributions compared to just point estimates; the distributions capture the uncertainty in the problem, presenting likely values of the unknowns that are different from simple point estimates.

Matched-field processing approaches are powerful tools for source localization and environmental parameter estimation in the ocean.Requiring multiple replica field calculations, however, matched-field processing can have significant computational demands. This work investigates the potential for using matched-field techniques that match only select "features" of the acoustic fields, attempting to reduce the computational requirements for successful inversion. The features this paper focuses on are arrival times of select paths. A "matching" technique for inversion using only arrival times is discussed and results are shown. A new process for the efficient selection of arrival times is also proposed.

Matched field processing (MFP) for geoacoustic inversion has been applied largely in the frequency domain using an incoherent combination of correlations at several frequencies. In source detection and localization problems, similar incoherent broadband MFP techniques have been shown to be inferior to coherent MFP techniques. Coherent MFP exploits source spectrum information when this is available, yielding more robust detection and localization results. This experience suggests that there is potential to improve geoacoustic inversion by MFP using a coherent algorithm. This work assesses that potential through a comparison of inversion results from incoherent and coherent methods, evaluating performance and cost. Estimates are made of water column depth and sediment geoacoustic properties for both synthetic data and real data from the SWellEX 96 experiment. The SWellEX data processed in this work correspond to source transmissions of lfm and hfm pulses, the spectra of which are employed in the coherent processing. [Work supported by ONR, Ocean Acoustics.]

No abstract

Estimation of time delays and amplitudes of multipath arrivals is of great interest in many fields because of the information that can be extracted from these characteristics. In underwater acoustics, multipath features of sound signals can reveal the location of the sound-generating source and properties of the propagating medium; their accurate estimation is, thus, desirable. In this work, a scheme is developed for the estimation of multipath arrival times and amplitudes in a noisy environment considering an unknown signal-to-noise ratio. The number of arrivals at the receiver is considered unknown as well, since, realistically, it depends on properties that would be uncertain a priori. The method is based on the efficient calculation of the number of paths, their amplitudes, and arrival times using Gibbs sampling to calculate a posteriori probability distributions. Results from the new method are compared to those of conventional approaches and show the potential of the new technique for efficient time delay and amplitude estimation in an environment with several unknowns. [Work supported by ONR.]