885095961

position parameters as measured by a hydrophone array, it will give the least achievable mean-square errors of source position. As is also well known, cstimation accuracy of a sourcc*s angular position rclative to an array is closely conncctcd to thc array’s apcrturc transvcrse to the linc of propagation between source and array. By application of the classical optics limit of the Cramer-Rao bound [A. B. Baggeroer et a!., J. A co ust. Soc. Am. 83, 571-587 (1988)] it becomes possible to transform a com-puted lower bound on angular uncertainty into an equivalent array apcrturc. It has recently bccn dcmonstrated [J. S. Perkms and W. A. Kuperman, J. Acoust. Soc. Am. 87. 1553-1556 (1990)] that a vertical array deployed in an azimuthally asymmetric environment will achieve angular resolution. Hence, the quantification of this relation between the environment and equivalent horizontal aperture via Cramer-Rao bounds is desired. Numerical modeling results indicating a vertical ar-ray’s equivalent horizontal aperture will be presented.

10:30

8UVV8. Application of sound to underwater acoustics. R. Pitre, Timothy L. Krout, John S. Perkins, B. Edward McDonald, W. A. Kuperman, Michael D. Collins, Nicholas C. Makris (Navat Res. Lab., Washington, DC 20375), and G. L. Gtbian (Planning Systems Inc., McLcan, VA 22102)

Experienccd sonar operators can rccognize various types of sounds. With the recent developement of time-domain underwater acoustics models and Computer workstations capable of generating high-quality sounds from digital data, it is now possible to train the human ear to associate a wide variely of sounds with the underlying physics. With sufhcient training, it might be possible to perform qualitative invcrsion problems such as estimating the rangę of a sound source, determining properties of the ocean sediment from a bottom-reflected signal. and recognizing the source of a scattered signal. Digitally generated sound can also be used in connection with time-domain signal-processing al-gorithms such as the delay-and-sum beamformer and the simulated annealing optimal beamformer [W. A. Kuperman et al._t J. Acoust. Soc. Am. 88, 1802-1810 (1990)]. The synthesized sounds presented will include propagation to various ranges in different ocean environments, refteclion from different types of ocean bottoms, scattering from bubble clouds near the ocean surface, and the application of beamforming to eztract signals from interference (including a single speaker in a noisy crowd).

10:45

8UW9. Multiple source behavior of passive synthetic aperture algorithms. Geoffrey S. Edclson (Dept. of Elcc. Eng., Univ. of Rhode Island, Kingston, RI 02881) and Edmund J. Sullivan (Codę 103, Naval Underwater Syst. Ctr., Newport, RI 02841)

Generally speaking, passive synthetic aperture algorithms involve the calculation of a single phase correclion factor from snapshot to snapshot in order to compensate for the spatial movement of the array over time. Within a given frequency bin and for sources in thc far field, this approach is shown to be suflficient for the case of any single source and for the case of two completely deterministic (fully coherent) sources. However, for any other scenario, such as one deterministic and one stochastic source, this techniquc is shown not to result in the desired phase corrcction factor. This is a direct consequcnce of the algorithms used in the various passive synthetic aperture tcchniques. Aftcr reviewing these tcchniques, an approach to the resolution of this problem, which is based on the multivariate maximum likelihood method, is presented.

11:00

8UW10. Channcl-adaptive matched filter Processing of large time-bandwidth product signals: Preliminary results. Jean-Pierrc Hermand (SACLANT Undersca Res. Ctr., APO, New York, NY

2001 J. Acoust. Soc. Am., Vol. 89. No. 4. Pt. 2. April 1991 09019) and William I. Roderick (Naval Underwater Syst. Ctr., New London, CT 06320)

Techniques that incorporate propagation channel modeling into conventional matched filter processing can improve dctection and local-ization performance of low-frequency active sonar Systems. In January 1989, an acoustic transmission experiment under known oceanographic conditions was conductcd by SACLANTCEN in a deep water (2800-m) area west of Sardinia. Large time-bandwidth product signals were transmitted by a broadband projector and were received at a 30-km rangę in the broadside bcam of a horizontal linę array of 64 hydrophones. Source and recciver were towed near the channel axis at depths of 100 and 150 m, respectively. The transmitted signals consisted of linear FM puLscs with time-bandwidth products that ranged from 400 to 4000. Environmental parameters, calculatcd from measured and archival data, were introduced in the GENERIC sonar model to com-pute the eigenrays at discrete frequencies closely spaced over the transmission bandwidth. The (band-limited) impulse response of the channel was obtained by complex summation of the eigenrays in both time and frequency domains. For each bandwidth, the transmitted signal, con-volved with the modeled channel response, was correlated against its replica. The predicted matched filter outputs were compared with those obtained from the obscrvations. The resulls will be discussed in the context of channel-adaptive matched filtering.

11:15

8UW11. Use of bispectral analysis in flow noise measurements. Mary F. Lei boli and Theordore M. Farabee (Codę 1942, Ship Acoust. Dept., David Taylor Res. Ctr., Bethesda, MD 20084)

Charactcrization of the wali pressure field in a turbulcnt boundary layer is necessary for accuratc models to be developed. Measurements of the wali pressure field of a turbulent boundary layer were madę in a quiet wind tunnel with flush-inounted pressure sensors. Bispectral analysis was performed on the resulting digitized data to determine whether frequency phase coupling was present. Bispectral analysis, which is an extcnsion of the commonly used Fourier analysis tcchniques known as auto- and cross-spectral analysis, identifies three wave phase couplings between frequencies by calculating thc expected value of the sum of the phases of three frequencies of interest:

= E{ |A'(/,)||Ar(/_J)||Y(/_l,_I)|

X exp[0(/,> + 0(/₂) + fl(/, , ₂)]}.

wherc | X{J) | and 0{f) are the rcspective magnitude and phase of the Fourier transform of *(/). Phase coupling can be indicative of nonlincar relationships in a system. [Work supporled by DTRC IR/IED Program]

11:30

8UW12. Evaluution of interference efTect in the estimation of underwater target abundance by the echo-intcgration method. Zhigang Sun and Gerard Gimcncz (Lab. “Traitement du Signal et Ultrasons," U.R.A. C.N.R.S. 1216, 1NSA, Bat. 502, 69621 Villeurbanne Ccdcx, France)

The estimation of fish abundance by the echo-integration method is based on the assumption that the total integrated echo intensity re-turncd from randomly distributed targets is proportional to their quan-tities. In the case where the multiple scattering between targets is neg-ligible, the precision of this estimation can be affected by the interference between individual cchos from cvery single target. In this work, an evaluation lunction to determine the eventual divergence of thc estimated result from the lincarity is proposed. As a function of different parameters, such as thc total number and the dimension of targets, the transmitted signal frequency, distancc from transduccr to the network plan and the transduccr bcam wid th, the importancc of interference effcct is analyzed for a network of spheres with various spatial distributions [Work supported by IFREMER.]

121st Meeting: Acoustical Society of America 2001