Real Acoustic Data Research Articles

Model selection of random amplitude polynomial phase signals

Procedures for selecting the amplitude and phase models of a polynomial phase signal (PPS) modulated by a nonstationary random process are presented. The model selection is performed by using nonlinear least-squares type phase and amplitude parameter estimators with sequential generalizations of the Bonferroni test. Simulation results are used to analyze the performance of the procedures. An application to modeling of real passive acoustic data is presented.

Aircraft flight parameter estimation using acoustical Lloyd's mirror effect

A model is developed for the acoustical Lloyd's mirror effect observed in the output time-frequency distribution of a microphone located near the ground during the transit of a jet aircraft. The feasibility of using this effect for flight parameter estimation is assessed by a simple Cramer-Rao lower bound analysis. The nonlinear least-squares method and the generalized Hough transform method are formulated for flight parameter estimation. The performances of both methods are evaluated and compared using real acoustic data.

A general procedure for estimating the composition of fish school clusters using standard acoustic survey data

An algorithm to identify classes of fish in acoustic backscatter images would improve the accuracy of acoustic biomass estimates over manually scrutinized images. A generalized Bayesian procedure for such identification called BASCET is presented, and two implementation strategies for the procedure are compared using simulated acoustic survey data. The procedure has several unusual characteristics: it evaluates schools not individually but in clusters; it makes use of human experience at cluster identification; it presents measures of uncertainty in all estimation results; and it constructs the training set required for supervised learning automatically using spatial and temporal assumptions. The simulation study comparison suggests that making use of temporal and spatial structure in the acoustic data leads to improved estimation performance. On the simulated data, the BASCET algorithm correctly identified the dominant fish class in 15 of 16 cases. However, the simulation model generates acoustic survey data based on the same assumptions used in BASCET, assumptions that may differ from a real acoustic survey. The study also assumed that the human experience incorporated in the Bayesian prior distributions was not misleading. Performance of BASCET on real acoustic data is presented in a companion paper.

Acoustic inversion via linearization and Bayesian multipath identification

A linearization approach to acoustic inversion is proposed employing distinct ray paths (direct arrival, first surface bounce, and first bottom bounce) for source localization and bathymetry and sound speed estimation. The ray path arrivals are selected from broadband, shallow water, synthetic data using a Bayesian time delay estimation scheme calculating posterior probability density functions of the delays in an efficient way. A linear system is then formed relating unknown parameters and arrival time data. The regularization method is used for the solution of the linear system [S. E. Dosso et al., J. Acoust. Soc. Am. 104, 846–859 (1998)] with excellent results. Also a simple least-squares approach for the solution of the system is implemented; results of the two approaches are compared. Finally, the linearization multipath based technique is successfully applied to real acoustic broadband data for source and receiver localization, and bathymetry and sound speed estimation. [Work supported by ONR.]

Signal processing for acoustic surveillance of the land environment

Surveillance systems that rely on the propagation of sound in the atmosphere are not constrained by electromagnetic line of sight and offer the advantages of being passive, lightweight, compact and covert. Such systems are unattended in their operation and are often deployed in remote areas. To demonstrate the scientific principles that underpin the operation of such systems, various narrow-band and broadband signal processing algorithms are applied to real acoustic data from both single and multiple sensors. In one example, the acoustical Doppler effect enables the flight parameters and blade-passage frequencies of turboprop and rotary wing aircraft to be estimated from the time-frequency signal analysis of spectrogram data. In another example, the acoustical Lloyd’s mirror effect is used to extract the flight parameters of a jet aircraft. Other examples deal with the application of generalized cross-correlation processing to the detection, localization and tracking of aircraft, ground vehicles and weapon fire. Wideband cross-correlation with differential Doppler compensation is required for the acoustic tracking of high-speed sources when the sensors are widely separated and phase tranform prefiltering is required to suppress the ambigouous peaks in cross-correlograms caused by the presence of strong narrow-band lines in the source spectra of air and ground vehicles.

Flight path estimation using frequency measurements from a wide aperture acoustic array

A narrowband technique based on the acoustical Doppler effect is proposed for estimating the trajectory of a turbo-prop aircraft in level flight with constant velocity as it transits over a ground-based passive acoustic sensor array. The basic principle is to measure the temporal variation of the instantaneous frequency (IF) of the acoustic signal received by each sensor and then to minimize the sum of the squared deviations of the IF estimates from their predicted values over a sufficiently long period of time for all sensors. The technique provides estimates of the propeller blade rate and the five source motion parameters that describe the aircraft trajectory. The six dimensional minimization problem is reduced to a five dimensional maximization problem, which is solved numerically using the quasi-Newton method. A simple method is described that provides the initial parameter estimates required for the numerical maximization. The effectiveness of the motion parameter estimation technique is verified using real acoustic data recorded from a wide aperture microphone array during various transits of a turbo-prop aircraft.

Broadband passive acoustic technique for target motion parameter estimation

A nonlinear least-squares method is formulated to estimate the motion parameters of a target whose broadband acoustic energy emissions are received by a ground-based array of sensors. This passive technique is applied to real acoustic sensor data recorded during the passage of a variety of ground vehicles past a planar cross array and its effectiveness verified by comparing the results with the actual values of the target motion parameters. The technique cam also be applied to airborne targets.

Passive wideband cross correlation with differential Doppler compensation using the continuous wavelet transform

An estimate of the time difference for the signal emitted by a stationary source to arrive at two spatially separated sensors is given by the time displacement that maximizes the cross-correlation function. For a fast moving source, however, this estimate is found to be in error because the time scales of the received signals are different for the two sensors. The correct time delay can be extracted by evaluating the continuous wavelet transform, which has the same functional form as the wideband cross-ambiguity function. When the signal-to-noise ratio is high, the coordinates of the ambiguity surface’s global maximum provide reliable estimates of both the differential time of arrival (or time delay) and the ratio of the time scales of the signals received by the two sensors. The continuous wavelet transform is computed using the one-step chirp z-transform method, the cross-wavelet transform method, and the two-step methods where multirate sampled replicas of the sensor waveforms are cross correlated, or else the sensor waveforms are interpolated using the discrete Fourier transform prior to cross correlation. The latter method is applied to real acoustic data recorded from an orthogonal configuration of three microphones during the low-altitude transit of a jet aircraft. The resulting time delay estimates are used to calculate the variation with time of the azimuth and elevation angles of the aircraft during the transit.

Application of the continuous wavelet transform to the angular location of jet aircraft

For a fast moving source and widely spaced sensors, the passive wideband cross-correlation method for time delay estimation requires compensation for the relative time scaling between the received signals, otherwise the time delay estimates will be in error. Passive wideband cross correlation with differential Doppler compensation is equivalent to evaluating the continuous wavelet transform, which is an operation that transforms a function (represented by the output waveform of one sensor) by integrating it with time-scaled and time-delayed versions of some kernel function (represented by the output waveform of the other sensor). When the signal-to-noise ratio is high, the coordinates of the wavelet transformed global maximum provide reliable estimates of both the time delay and the ratio of the time scales of the signals received by the two sensors. The wavelet transform is used to process real acoustic data recorded from an orthogonal configuration of three microphones during the low altitude transit of a jet aircraft. The resulting time delay estimates are used to calculate the variation with time of the azimuth and elevation angles of the aircraft during its transit.

A robust adaptive beamformer for microphone arrays with a blocking matrix using constrained adaptive filters

This paper proposes a new robust adaptive beamformer applicable to microphone arrays. The proposed beamformer is a generalized sidelobe canceller (GSC) with a new adaptive blocking matrix using coefficient-constrained adaptive filters (CCAFs) and a multiple-input canceller with norm-constrained adaptive filters (NCAFs). The CCAFs minimize leakage of the target-signal into the interference path of the GSC. Each coefficient of the CCAFs is constrained to avoid mistracking. The input signal to all the CCAFs is the output of a fixed beamformer. In the multiple-input canceller, the NCAFs prevent undesirable target-signal cancellation when the target-signal minimization at the blocking matrix is incomplete. The proposed beamformer is shown to be robust to target-direction errors as large as 200 with almost no degradation in interference-reduction performance, and it can be implemented with several microphones. The maximum allowable target-direction error can be specified by the user. Simulated anechoic experiments demonstrate that the proposed beamformer cancels interference by over 30 dB. Simulation with real acoustic data captured in a room with 0.3-s reverberation time shows that the noise is suppressed by 19 dB. In subjective evaluation, the proposed beamformer obtains 3.8 on a five-point mean opinion score scale, which is 1.0 point higher than the conventional robust beamformer.

Towed-array beamforming during ship's manoeuvring

Towed hydrophone arrays are commonly used for determining the spatial characteristics of the underwater acoustic field. The assumption that the hydrophones lie in a straight and horizontal line is often made when beamforming the hydrophone outputs. However, due to tow vessel motion, ocean swells and currents the array adopts a nonlinear shape and the beamformer output is degraded. To estimate the positions of the hydrophones an array was instrumented with a set of positioning sensors: compasses, tiltmeters, accelerometers and pressure gauges. The authors present the array deformations recorded at sea when the tow vessel is turning and along straight-line tracks. The influence of the observed deformations on the performance of the conventional beamformer output is discussed and illustrated with simulated and real acoustic data.

Application of a maximum likelihood processor to acoustic backscatter for the estimation of seafloor roughness parameters

Maximum likelihood (ML) estimation is used to extract seafloor roughness parameters from records of acoustic backscatter. The method relies on the Helmholtz–Kirchhoff approximation under the assumption of a power-law roughness spectrum and on the statistical modeling of bottom reverberation. The result is a globally optimum, highly automated technique that is a useful tool in the context of seafloor classification via remote acoustic sensing. The general geometry of the Sea Beam bathymetric system is incorporated into the design of the ML processor in order to make it applicable to real acoustic data collected by this system. The processor is initially tested on simulated backscatter data and is shown to be very effective in estimating the seafloor parameters of interest. The simulated data are also used to study the effect of data averaging and normalization in the absence of system calibration information. The same estimation procedure is applied to real data collected over two central North Pacific seamounts, Horizon Guyot and Magellan Rise. The Horizon Guyot results are very close to estimates obtained through a curve-fitting procedure presented by de Moustier and Alexandrou [J. Acoust. Soc. Am. 90, 522–531 (1991)]. In the case of Magellan Rise, discrepancies are observed between the results of ML estimation and curve fitting.

Beamforming towed array data when knowledge of the hydrophone positions is imperfect

Information on the various sound sources that contribute energy to the underwater acoustic environment is often obtained by processing the data from a thin flexible line array of hydrophones towed behind a vessel. The acoustic data from the array are processed digitally in the frequency and spatial domains. The process of spatial filtering, which is commonly referred to as beamforming, assumes that the horizontal line array is straight with the transverse displacements of the hydrophones being zero with respect to the array axis. Using real acoustic data, this paper shows the effects on conventional and adaptive beamformers when these transverse displacements are finite and the array shape is nonlinear. Furthermore, signal processing techniques are presented that significantly reduce the influence of these effects on the performance of the beamformers. By processing the array’s acoustic data when a plane-wave acoustic source is present, these techniques estimate the horizontal and transverse displacements of the hydrophones for subsequent input to the spatial filters.

Sharpness applied to the adaptive beamforming of acoustic data from a towed array of unknown shape

When a vessel towing a horizontal line array of underwater acoustic sensors changes course, the array responds by changing its shape so that the sensors no longer lie along a straight line. However, the spatial processing of the acoustic data still proceeds on the assumption that the array is straight with the positions of the sensors being known and invariant. When the array adopts a nonlinear shape so that the actual locations of the sensors no longer coincide with the assumed positions, the performance of the beamformer is observed to deteriorate. This degradation is most marked when an adaptive beamformer is used to optimize the array gain. The adaptive beamformer considered here is a constrained optimum beamformer that is based on the inversion of the observed (signal-plus-noise) cross-spectral matrix of the sensor outputs. The adaptive beamformer responds to any deviation of the sensors from their nominal positions by suppressing the received signals so that the output signal-to-noise ratio of the beamformer, and hence the array gain, decrease in the direction of each signal. One method of overcoming the array shape problem is to monitor the geometrical configuration of the array’s sensors by instrumenting the length of the array with compasses and depth sensors. This paper examines an alternate method whereby data from the acoustic sensors themselves are used to infer the shape of the array. This narrow-band acoustic technique, which relies on at least one acoustic source being present in the far field, evaluates a sharpness function for a variety of possible array shapes. The sharpness is calculated by summing the product of the beam output power squared and the sine of the beamsteer angle over all beamsteer directions from forward endfire to aft endfire. Using simulated data, it is shown that the estimated array shape matches the actual shape when the sharpness attains its maximum value. Also shown is the dramatic reduction in signal suppression when this acoustic technique is applied to the adaptive beamforming of real acoustic data.

Measurement of Directionality and Distribution in an Acoustic Field

The fact has been demonstrated previously, and is briefly reviewed here, that characteristic sounds occurring in an acoustic field can be identified by means of their first-order probability distributions. In the present paper, it is shown that still more identifying information becomes available when the directionality of the acoustic field is also measured, thus adding space processing to the previous time processing. The concept of directionality is defined quantitatively in this paper, signal processing principles, which have been derived for estimating the directionality statistic, are described, making use of the multimode hydrophone [S. L. Ehrlich and P. D. Frelich, U. S. Pat. 3,290,646, 1966 filed 6 (Apr. 1960)], and the results of playing real acoustic field data into a computer simulation of this directionality detector, together with the previously derived distribution detector, are presented. Detected in the acoustic data are a monochromatic plane wave, a rather directional transient, and nonisotropic Gaussian ambient noise.