LL Logo



Robust Adaptive Array
Processing in Stochastic
Shallow-Water Multipath
Acoustic Channels

Jeffrey L. Krolik
Duke University
Department of Electrical Engineering
Durham, NC 27708-0291
email: jk@ee.duke.edu

Passive array processing methods which use full-wave models of complex multipath propagation to facilitate range/depth localization of underwater acoustic sources are known to be extremely sensitive to errors in the assumed environmental conditions. This presentation addresses adaptive array processing methods which exploit statistical characterizations of uncertain environmental conditions to achieve improved robustness to signal model mismatch. In particular, methods are presented which use the second-order statistics of hypothesized source wavefronts in the design of linear constraints for a minimum variance adaptive beamformer. Using simulated stochastic ocean channel models derived from real environmental data, the probability of correct source localization achieved by the robust methods is shown to be significantly higher than for conventional techniques. Further, successful range/depth localization of a real acoustic source in the Mediterranean without precise knowledge of the environmental conditions is demonstrated using NATO SACLANT Center vertical array data. In an extension of this approach, techniques for using wavefront measurements from a source-of-opportunity to improve localization accuracy are also described. Assimilation of the source-of-opportunity wavefront estimates into design of the adaptive beamformer constraints is achieved by computing the conditional correlation matrix of hypothesized source wavefront vectors using a stochastic adiabatic normal mode model. The method also includes an accurate procedure for estimating a source-of-opportunity wavefront in a multiple source environment.

This work is supported by ONR.



LL Logo Disclaimer

Direct comments and questions to: webmaster@ll.mit.edu

MIT Lincoln Laboratory. All rights reserved.