885095960

erable variation in the sidelobes as a function of frequency while the main peak remains at the sourcc location. [Work supported by the Office of Naval Research.) the limiting behavior of the approximate likelihood-ratio statistic as the signal-to-noise ratio vanishes. [Work supported by the U.S. Office of Naval Research under Grant N00014-89-J-1321.]

9HK)

8UW3. Broadband modeling and source tocalization of a PRN projector tow. Evan K. Wcstwood (Appl. Res. Labs., Univ. of Texas at Austin. Austin, TX 78713-8029)

A broadband ray model is used to simulate data from a deep water experiment (TAGEX 87) in which a pseudo random noise (PRN) source is towed past a 24-elemcnt, bottom-moored array. The source spectrum contains a senes of evenly spaced lines from 60-100 Hz. Sim-ulated shade plots of received spectrum łeve! versus frequency and time exhibit excellent agreement with the data. A broadband matched field algorithm is used to localize the source in rangę as a function of time. The algorithm consists of matching the rccorded cross spectrum R\Rf between two phones with simulated cross spectra S_tSf at a senes of ranges. The correlations between recorded and simulated cross spectra are added coherently as a function of frequency to obtain the local-ization matrix entry: L,j — 2/lRuRf,]    Successful localizations of

the PRN source are obtained at ranges from 0-30 km. (Work supported by Naval Ocean Systems Center, Codę 733, under contract N00039-88-C-0043-1-43.)

9:15

8UW4. A new techniquc for detecting Gaussian signals in non-Gaussian noise. H. Vincent Poor (Dept. of Elect. Eng., Princeton Univ., Princeton, NJ 08544)

The signal and noise processes arising in passive underwater acous-tic detection are usually modcled as being random processes, with the dominant noise often being best modeled as a non-Gaussian process (H. V. Poor and J. B. Thomas, J. Acoust. Soc. Am. 63, 75-80 (1978)] due to the effeets of noise phenomena resulting from sources such as crack-ing ice, marinę animals, or surface shipping. A new technique for the detection of Gaussian signals that can be modeled as being produced by linear stochastic systems, in the presence of such non-Gaussian noise, has been developed. This technique is based on an approximation to the likelihood-ratio statistic (H. V. Poor, An Introduction to Signal Detection and Estimation (Springer-Verlag, New York, 1988)] for such situ-ations. This likelihood-ratio approximation is in tum based on the Mas-reliez approximation of nonlinear filtering [R. Vijayan and H. V. Poor, IEEE Trans. Commun. 38, 1060-1065 (1990)], in which the predicted State probability density (i.e., the time update) of the underlying stochastic system is approximated with a multivariate Gaussian distribu-tkm. This approximation allows calculation of the likelihood-ratio statistic using a pair of sufficient statistics satisfying a simple nonlinear recursion. An approximation to the locally optimum detection statistic [Poor, op. cit.) for this situation has also been derived, by considering

9:30

8UW5. Joint estimation of rangę and bearing in the presence of fluctuation. James F. Bart ram and Sudha S. Reese (Raytheon Co., Submarine Signal Division, 1847 W. Main Rd., Portsmouth, RI02871)

Active sonar joint range-bearing estimation is treated, based on the use of a signal with a large bandwidth-timc (WT) product, specifically one with linear frcquency modulation (LFM), together with a Processing scheme in which rangę is estimated by means of pulse compression and time-delay measurement, and bearing is estimated using split-array differential phase processing. The degradation in rangę resolution caused by fluctuation is derived using a simple channel spreading model, whose parameters are B, the Doppler spread, and L, the delay spread, while the degradation in bearing precision is evaluated on the assump-tion that the physical fluctuation process will cause the retuming signal to become normally distributed. It is shown that temporal segmentation of the waveform can have a beneficial effcct. The optimum number of segments to use is determined, depending on the WT product of the signal, on the one hand, and the BL product characteristic of the fluctuation process on the other. The cffect of using a diflerent typc of frcquency modulation in the signal design remains for futurę study.

9:45

8UW6. Modal decomposition of the pressure field on a vertical linę array. David J. Thomson, Gordon R. Ebbeson, and Brian H. Maranda (Defence Res. Establishment Pacific, FMO Victoria, British Columbia V0S IR0, Canada)

For many applications, it is useful to decompose the acoustic field me as u red in shallow water into its horizontal wave-number compo-nents. Recently, a PE-based method was described for cffccting the modal decomposition of the depth-dependent field at a given rangę in a range-dependent waveguide [D. J. Thomson, J. Acoust. Soc. Am. Suppl. 1 86, S53 (1989)]. This method should be applicabie to the analysis of data obtained with a vertical linę array (VLA). However, for practical arrays, the measured field is known at only a limited number of hydro-phones, whereas the PE-based decomposition method requires the field to be known at each depth on the computational grid. Therefore, to populatc the entire PE grid, it is necessary to interpolatc the field between hydrophones and to extrapolate it into the bottom. An interpo-lation and extrapolation scheme suitable for this purpose is proposed. To illustrate the effectiveness of this reconstruction scheme, modal decomposition is carried out using simulated VLA data generated for an envirunment representative of the Continental shelf region of the Cana-dian Arctic.

10:00-10:15

Break

10:15

8UW7. Cramer-Rao bound characterization of equivalent horizontal aperture achieved from environmental asymmetry. John Glattetre (Norwegian Defence Res. Establishment, P. O. Box 115, N-3191

Horten, Norway), John S. Pcrkins, and W. A. Kuperman (Naval Res. Lab., Washington, DC 20375-5000)

The Cramer-Rao bound is a well-accepted lower bound on the mean-square error of estimated values. Applied to estimation of source

2000

J. Acoust. Soc. Am., Vol. 89, No. 4, Pt. 2, April 1991

121st Meeting: Acoustical Society of America

2000