Simulation of Ocean Acoustic Tomography Using Matched Field Processing
Abstract
The feasibility of applying the principles of matched field processing to ocean acoustic tomography were studied under various conditions of ambient noise. Several likelihood estimators were examined (e.g., Bucker, Bartlett, Maximum Likelihood, etc.). Simulations were initially conducted for the simple case wherein only one parameter of the medium was unknown (e.g., SOFAR axis depth, surface sound speed, position of a single acoustic front). The method was then applied to the more realistic problem of locating the boundaries of an eddy in the ocean. For moderate signal-to-noise ratios, all the estimators were shown to be able to solve the problem, albeit with different efficiencies. For low signal-to-noise ratios, the MLM scheme proved to be the most reliable especially when a highly correlated ambient noise was present. In all cases, computer simulations illustrated that mismatching may occur when the parameterization of the medium is poorly approximated. Mismatching leads to a decrease in the efficiency of the estimators but it may be still possible to correctly estimate the environmental characteristics. Keywords: Theses.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1989
- Accession Number
- ADA216498
Entities
People
- Frederic Strohm
Organizations
- Naval Postgraduate School