Sunday, December 17, 2017
      CMRE Facebook page  CMRE LinkedIn page  CMRE PAO Youtube page
Text Size
CMRE banner

Formal Reports

Report of results of completed projects or major milestones either in scientific terms or in terms acceptable to a wider audience. Note: Unless linked to the full text, reports are only available to NATO member nations from designated distribution centres. 


Order by : Name | Date | Hits | [ Ascendant ]

Bayesian inference of seabed geoacoustic and scattering properties using the DREAMzs algorithm Bayesian inference of seabed geoacoustic and scattering properties using the DREAMzs algorithm

Date added: 10/07/2015
Date modified: 10/07/2015
Filesize: Unknown

Bayesian inference of seabed geoacoustic and scattering properties using the DREAMzs algorithm.  Nielsen, Peter L.; Canepa, Gaetano; Fox, Warren L. J. CMRE-FR-2015-001. January 2015.

The probability of mine burial and sonar performance prediction are strongly dependent on the seabed properties. In particular, for low-frequency mine-hunting sonar operations focusing on detection and classification of partial or fully buried targets, the transmitted acoustic field has significant interaction with the seabed. CMRE is in the process developing a sonar system operating at lower acoustic frequencies compared to more traditional mine-hunting sonars, and the intention is to use this sonar system to determine the seabed properties as the mine hunting mission progresses. This report describes a methodology to estimate the unknown seabed geoacoustic and scattering properties with associated uncertainties tailored to the CMRE low-frequency sonar. The algorithm is composed of an acoustic model developed at CMRE to calculate the backscattered intensity from a layered seabed as a function of geoacoustic, interface and volume scattering properties. This acoustic model is linked to an inversion package based on a Bayesian framework to provide the best estimate and uncertainties of the seabed properties. The performance of this environmental characterization algorithm is demonstrated for synthetically generated test cases which mimic the operation of the CMRE low-frequency mine-hunting sonar. The test cases are at present limited to a fine silt bottom type described as an infinite halfspace with water-seabed interface and volume scattering. Both noise free and noisy synthetically generated data are analyzed, and different run parameters for the inversion package are changed to evaluate the impact on the inferred seabed properties. The added noise is treated as an unknown together with the seabed properties. The geoacoustic and scattering parameters are well determined for the noise free cases, while only sediment sound speed, volume scattering spectral level and added noise are well estimated for the noisy test cases.

User Login