C. Soares and S.M. Jesus
csoares@ualg.pt sjesus@ualg.pt
SiPLAB - FCT, Universidade do Algarve,
Campus de Gambelas,
PT-8005-139 Faro, Portugal.
E. Coelho coelho@nrlssc.navy.mil
Emanuel F. Coelho
Naval Research Laboratory
Code 7322, Bldg 1009, Room C128
Stennis Space Center, MS39529
Abstract
Environmental inversion of acoustic signals for seafloor and water column properties
has been proposed as a concept for complementing direct oceanographic measurements for
Rapid Environmental Assessment (REA). In order to respond to operationality requirements
of acoustic REA a light receiver device, an Acoustic Oceanographic Buoy (AOB)
with a sparse vertical array, was developed under the AOB-REA Joint Research Project.
Inversion methods based on Matched-Field Processing appear to be suitable to be applied
under those requirements. However, if a reduced number of receivers is available to sample
the vertical normal modes of the acoustic field, then the inverse problem is heavily
ill-conditioned, whereas the underlying cost-function may have a massive sidelobe structure
with many local extrema. This causes difficulties to meta-heuristic search methods, such as
a genetic algorithm, 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, under the belief that this can contribute for improving convergence. A
minimum-variance and a subspace based method are applied to the inversion of synthetic
and experimental acoustic data using a sparse array, and compared to a Bartlett processor.
Their effectiveness in multi-parameter estimation problems is clearly demonstrated.
Acknowledgement: The authors would like to thank the NATO Undersea Research Centre for the organization of the MREA'03 sea trial. This work was financed by FCT, Portugal, under fellowship SFRH/BD/12656/2003 and NUACE project, contract POSI/CPS/47824/2002, and the Portuguese Ministery of Defense under the LOCAPASS project.