LL Logo



Adaptive Beamforming
(ABF) and Direction
Finding (DF) for Arrays in
Difficult Environments

Norman L. Owsley
Naval Undersea Warfare Center (NUWC)
New London, CT 06320
email: owsleyl@nl.nuwc.navy.mil

Abstract Acoustic energy in a shallow water, bottom limited environment is complicated by multiple boundary (surface and bottom) interactions and severely refracting sound velocity profiles. Attempts to model and/or estimate the point-to-point, full-field transfer functions necessary to perform environmentally matched ABF in such an environment have been, at best, marginally successful and academically stimulating at high SNRs and, at worst, marginally useless at low SNR in the presence of uncertainty in the propagation model. In this presentation a hierarchy of environmental model mismatch conditions are considered including: 1) simple beam steering errors; 2) coherent multipath (-mode) propagation model errors; and 3) azimuth angle spreading model errors.

It is observed that in a difficult environment the detection problem is arguably blessed with low SNR. The downside to the low SNR syndrome is that it does not provide a realistic opportunity for some form of an adaptive matched field processor to learn "everything you always wanted to know about the ocean'' propagation model using energy propagating from a source which is being detected simultaneous with environmental learning. The upside of the low SNR syndrome is that types (1) and (2) mismatch become asymptotically unimportant at low SNR. That is to say, low SNR performance with mismatch can be no worse than a conventional single wavefront time delay-and-sum beamformer which is notably robust to propagation model errors (mismatch) for all SNR. This may be a widely known but insufficiently understood result.

For type (3) mismatch, the issue is that in difficult environments the spatial coherence of the signal wavefront is not maintained over the horizontal aperture of the receiving array. A commonly proposed approach to this condition is the incoherent recombination of coherently processed horizontal subapertures of the full array. It is argued that distributing the sensors volumetrically over a reduced horizontal aperture with fully coherent array processing is preferred.

Actual sea trial data is presented which illustrates most of the important points of the presentation.



LL Logo Disclaimer

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

MIT Lincoln Laboratory. All rights reserved.