Acoustic Vector Sensor Research Articles

Determination of the location of a sound source in three dimensional based on acoustic vector sensors on the ground.

An acoustic vector sensor (AVS) consists of three orthogonal particle velocity sensors in combination with a sound pressure microphone. In several publications it has been proven that multiple sources can be located in three dimensions with a single AVS. In this paper it will be shown that it is possible to measure the instantaneous location (this means bearing, elevation, and range) of a single dominant sound source in three dimensional space as well as the angle dependent local ground impedance. Theory as well as results of experiments will be presented.

An ultra-low-frequency acoustic vector sensor.

Passive underwater acoustic surveillance and oceanographic studies, for the most, have not been conducted at ultra-low frequencies (ULFs), from 0.01 Hz to a few hertz. One of the primary reasons for neglecting ULF has been the lack of sensitive, and directional, ULF acoustic sensors. Existing, and many prototyped, vector sensor designs are inertia based and thus not well suited to ULF, typically these sensors that operate at frequencies well above 100 Hz. A new type of acoustic vector sensor, capable of operating within the ULF range and possessing unsurpassed sensitivity (minimal detectable signal levels well below 1 nm/s) was built, tested, and calibrated at the Naval Undersea Warfare Center, Newport, RI. The measurements have validated the proposed detection concept, provided initial performance parameters, and have set forth the foundation for designing the next generation of fieldable prototypes.

Amplitude, phase, location, and orientation calibration of an acoustic vector sensor array. Part II. Experiments.

An acoustic vector sensor array consists of multiple sound pressure microphones and particle velocity sensors. A pressure microphone usually has an omni-directional response, yet a particle velocity sensor is directional and usually has a response pattern as a figure of number eight. Currently, acoustic vector sensor arrays are under investigation for far field source localization and visualization. One of the major practical issues in these applications, however, is to determine the accurate position, orientation, and complex (phase and amplitude) sensitivity of each sensor within the array. In this study, a calibration method, which is discussed in Part I, is verified with experiments. The method determines all the crucial parameters based on a limited number of measurements with a reference sensor and multiple sound sources located at known locations. The calibration method shows also to be robust to mistakenly switched cable connections. The experiments are performed in different acoustic environments including an anechoic room. The results will be presented in this paper.

Three dimensional sound source localization and sound mapping using a p-u sensor array.

An acoustic sensor array, which consists of pressure and particle velocity sensors, is an attractive alternative to phased pressure array because knowledge of the three dimensional (3-D) particle velocity directly characterizes the direction of the source. Hence, it is possible to localize sources in the low-frequency range, where the phase difference between pressure sensors is small, and in the high-frequency range, where aliasing occurs with pressure sensors. This article applies advanced source localization techniques from aeroacoustics to acoustic vector sensors. Several methods are simulated and tested with several configurations and validated using measurement data from an anechoic chamber. The setup to test the algorithms consists of 24 sources that are distributed in 3-D around the set ups in the anechoic chamber. The sources are uncorrelated driven with white noise. Conventional beamforming methods have been compared with acoustic vector based deconvolution methods and MUSIC algorithms. Four configurations are tested: (1) a planar array of sound pressure microphones, (2) a planar array of pu probes (pu probes are collocated sound pressure and particle velocity probes); the velocity probes where the velocity probes are oriented in randomly, (3) a velocity gradient array, (4) a quad array of three dimensional velocity probes plus collocated sound pressure. The different methods and algorithms

Amplitude, phase, location, and orientation calibration of an acoustic vector sensor array, Part I: Theory.

An acoustic vector sensor array consists of multiple sound pressure microphones and particle velocity sensors. A pressure microphone usually has an omni-directional response, yet a particle velocity sensor is directional. Currently, acoustic vector sensor arrays are under investigation for far field source localization and visualization. One of the major practical issues in these applications, however, is to determine the accurate position, orientation, and complex (phase and amplitude) sensitivity of each sensor within the array. In this study, a calibration method is developed to determine each of those crucial parameters based on a limited number of measurements with a reference sensor and multiple sound sources located at known locations. The calibration method is also designed to be robust to mistakenly switched cable connections. Ideally, the calibration process should take place in an anechoic environment, but efforts have been made to compensate the effects of moderate background noise and reflections.

Acoustic Vector-Sensor FFH "Blind" Beamforming & Geolocation

An acoustic direction-finding and blind interference rejection algorithm for up to three noncooperative wideband, fast frequency-hop (FFH) spread spectrum signals of unknown hop sequences and unknown arrival angles is presented. This algorithm avoids the need to adjust beamforming weights to accommodate frequency variations in the frequency-hopping signals. This approach deploys one acoustic vector-sensor (a.k.a. vector-hydrophone), which is composed of two or three spatially collocated, but orthogonally oriented, acoustic velocity-sensors plus an optional collocated pressure-sensor. The spatial collocation of all component sensors comprising such an acoustic vector-sensor frees the beamforming weights of any dependence on the signals' hopping frequencies.

Geoacoustic inversion using specific acoustic impedance.

An inversion method for the estimation of geoacoustic properties of the ocean bottom using data from a small number of acoustic vector sensors was developed. The experimental data used for the inversions were short duration, gated continuous wave transmissions acquired on four acoustic vector sensors that spanned the water-sediment interface during the Sediment Acoustics Experiment 2004 (SAX04) conducted in the Gulf of Mexico. The acoustic signals were analyzed for their specific acoustic impedance over a bandwidth ranging from 600 to 2400 Hz. The specific acoustic impedance was obtained from simultaneous measurements of the acoustic pressure and particle acceleration for arrivals at normal incidence. A differential evolution algorithm was used for the inversion of the depth dependent sediment properties. A wave number integration routine was used to compute the specific acoustic impedance corresponding to the estimated sediment parameters. Experiment details, forward modeling, inversion methods, and results are discussed. [Work supported by ONR.]

CramÉr-Rao Bounds for Direction Finding by an Acoustic Vector Sensor Under Nonideal Gain-Phase Responses, Noncollocation, or Nonorthogonal Orientation

An acoustic <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">vector</i> -sensor (also known as vector-hydrophone in underwater applications) is composed of two or three spatially collocated but orthogonally oriented acoustic <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">velocity</i> sensors, plus possibly a collocated acoustic <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">pressure</i> sensor. Such an acoustic vector sensor is versatile for direction-finding, due to its azimuth-elevation spatial response's independence from the incident source's frequency, and bandwidth. However, previously unavailable in the open literature is how the acoustic vector sensor's far-field direction-of-arrival estimates may be adversely affected by any unknown nonideality in the acoustic vector sensor's gain response, phase response, collocation, or orthogonal orientation among its constituent velocity sensors. This paper pioneers a characterization of how these various unknown nonidealities degrade direction-finding accuracy, via Cramer-Rao bound analysis.

Subspace intersection method of high-resolution bearing estimation in shallow ocean using acoustic vector sensors

The subspace intersection method (SIM) provides unbiased bearing estimates of multiple acoustic sources in a range-independent shallow ocean using a one-dimensional search without prior knowledge of source ranges and depths. The original formulation of this method is based on deployment of a horizontal linear array of hydrophones which measure acoustic pressure. In this paper, we extend SIM to an array of acoustic vector sensors which measure pressure as well as all components of particle velocity. Use of vector sensors reduces the minimum number of sensors required by a factor of 4, and also eliminates the constraint that the intersensor spacing should not exceed half wavelength. The additional information provided by the vector sensors leads to performance enhancement in the form of lower estimation error and higher resolution.

Robust adaptive acoustic vector sensor beamforming using automated diagonal loading

In this paper, a novel robust adaptive acoustic vector sensor beamformer based on shrinkage is derived. Unlike many existing methods, the proposed method is completely automatic (or so-called user parameter-free), which means, it do not need the choice of user parameters. The proposed diagonal loading algorithms use shrinkage-based covariance matrix estimates, instead of the conventional sample covariance matrix, in the standard Capon acoustic vector sensor beamforming formulation. The numerical results show that our method is robust against errors on the steering vector and small sample sizes, and meanwhile gives high output signal to interference plus noise ratio (SINR).

Performance analysis of velocity–pressure smoothing for direction‐of‐arrival estimation of coherent signals

AbstractIn this paper, we consider the problem of direction‐of‐arrival (DOA) estimation of coherent signals by using a number of identically oriented acoustic vector‐sensors, wherein source decorrelation can be achieved by velocity–pressure smoothing. We examine the effect of velocity–pressure smoothing on the performance of MUSIC‐type methods in terms of the overall root mean‐square errors (RMSE) of DOA estimates. The closed‐form expression for the overall RMSE is derived and analyzed. We also compare velocity–pressure smoothing with the conventional spatial smoothing technique. Simulation results are given to confirm the analyses. Copyright © 2009 John Wiley &amp; Sons, Ltd.

On the use of acoustic particle motion in geoacoustic inversion.

Geoacoustic inversion estimates sediment properties based on one or more parameters of an observed acoustic field via inverse mathematical methods. The observed acoustic parameter is usually derived from the acoustic pressure measured at one or more locations. Recent advances in acoustic sensor technology have enabled the simultaneous measurement of the acoustic pressure and particle motion in three dimensions. The additional information provided by such acoustic vector sensors can be used to improve the performance of existing and novel geoacoustic inversion techniques. Current research seeks to use the additional information that is provided by the acoustic vector sensor to pose new inverse problems for the estimation of seabed sediment properties. In particular, data collected during the Sediment Acoustics Experiment 2004 (SAX04) included acoustic pressure and particle acceleration from a small number of acoustic vector sensors arranged in a vertical line, spanning the water‐sediment interface. A varie...

Direction finding in spatially correlated noise fields with arbitrarily-spaced and far-separated subarrays at unknown locations

A computationally simple azimuth-elevation direction finding algorithm in spatially correlated noise fields using two-far-separated subarray geometry (Li et al. 1995) is presented. We assume one subarray consists of multiple acoustic vector sensors, while another subarray comprises multiple pressure sensors. All sensors are arbitrarily placed at unknown locations. The authors firstly define a cross matrix to eliminate the effect of the spatially correlated noise. Then the so-called propagator method is used to estimate the steering vectors of acoustic vector sensors. Finally, a closed form, automatically paired azimuth-elevation angle estimates are derived. The proposed algorithm requires no eigen decomposition into signal or noise subspaces. In addition, the proposed algorithm does not need 2D iterative searching. Therefore the algorithm shows low computational complexity. Monte Carlo simulations are presented to verify the effectiveness of the proposed algorithm.

Effects of shear waves on boundary-coupled vector sensors

The impact of the acoustic properties of a solid medium on the response of an acoustic vector sensor placed at the boundary between water and the solid is examined. In particular, the components of the vector field both normal and tangential to the boundary are studied. It is shown that the component of the velocity tangential to the boundary is affected by the introduction of shear in the solid medium, which leads to amplitude and phase discontinuities across the boundary. Such discontinuities can lead to mismatches between the vector sensor components that interfere with common, coherent processing techniques. Implications for ocean bottom- and hull-mounted examples are provided.

Fast direction of arrival algorithm based on vector-sensor arrays using wideband sources

An acoustic vector sensor (AVS) can capture more information than a conventional acoustic pressure sensor (APS). As a result, more output channels are required when multiple AVS are formed into arrays, making processing the data stream computationally intense. This paper proposes a new algorithm based on the propagator method for wideband coherent sources that eliminates eigen-decomposition in order to reduce the computational burden. Data from simulations and lake trials showed that the new algorithm is valid: it resolves coherent sources, breaks left/right ambiguity, and allows inter element spacing to exceed a half-wavelength.

Two-dimensional direction finding of acoustic sources by a vector sensor array using the propagator method

This paper considers a new azimuth-elevation direction finding algorithm for multiple acoustic plane wave signals using acoustic vector array. Firstly, a planar-plus-an-isolated sensor array geometry [P. Li, J. Sun, B. Yu, Two-dimensional spatial spectrum estimation of coherent signals without spatial smoothing and eigendecomposition, IEE Proc.-Radar Sonar Navigat. 143(5) (October 1996) 295–299] is exploited and a cross-covariance matrix is defined. Then the propagator method is used to estimate the steering vectors of acoustic vector sensors. Finally, a closed-form, automatically paired azimuth-elevation angle estimates are derived. The presented algorithm shows high azimuth-elevation estimation accuracy due to array aperture extension. In addition, the new algorithm does not need the eigen-decomposition and 2D iterative searching, and is applicable to coherent (fully correlated) signals and spatially correlated noises. Therefore, the algorithm shows low computational complexity and robustness. Monte-Carlo simulations are presented to verify the effectiveness of the proposed algorithm.

An autonomous moored ocean profiler

An unconventional set of environmental and acoustic data, measured from an autonomous moored ocean profiler, is presented. The profiler is a programmable robotic underwater winch and data acquisition system that cycles vertically through the water column. The system profiles at pre‐determined time intervals, surfaces to transmit data via RF links, and can be deployed for up to 180 days. Built by WETLabs Incorporated (Western Environmental Technology Laboratories) for the Naval Underwater Systems Center (NUWC), the system is outfitted with three environmental sensors and two acoustic vector sensors. The environmental sensors include; a Nortek Vector velocimeter which samples current components and characterizes surface wave velocities, a Seabird FastCAT Conductivity, Temperature, Depth (CTD) sensor; and a Wetlabs Fluorometer‐Turbidity optical sensor that measures both chlorophyll‐a and turbidity. The Wilcoxon acoustic vector sensors measure acoustic pressure and the three components of acoustic particle acceleration. Wilcoxon Research (a subsidiary of Meggitt Corporation) and Applied Physical Sciences (APS) have developed underwater acoustic vector sensors for many applications. The quality and sensitivity of these vector sensors has steadily improved. A review of early (pre‐2002) vector sensor development efforts is discussed.

Acoustic sensor systems on a flying wing underwater glider and two prop‐driven autonomous underwater vehicles

The Marine Physical Laboratory, Scripps Institution of Oceanography operates several underwater vehicles including an autonomous underwater glider based on a flying wing design and two prop‐driven autonomous underwater vehicles (AUV) manufactured by Bluefin Robotics. The objective of this presentation is to describe the acoustic sensor systems on these platforms and provide sample results from the at‐sea data. The glider, with a 6.1‐m wing span, was developed jointly by the Marine Physical Lab and the Applied Physics Laboratory, University of Washington. It is designed to maximize the horizontal distance traveled between changes in buoyancy (i.e., maximize its "finesse") while quietly listening to sounds in the ocean. A 27‐element hydrophone array with 5 kHz per channel bandwidth is located in the sonar dome all along the wing's leading edge. In addition, it has a four‐component acoustic vector sensor in its nose. The two prop‐driven AUVs have been equipped with hull‐mounted hydrophone arrays with 10 kHz bandwidth for passive synthetic aperture studies, an acoustic vector sensor, and active acoustic imaging systems for ocean bottom/subbottom mapping. Results from the data processing illustrate the tight coupling between acoustic sensor systems, signal/array processing methods, and vehicle behavior. [Work supported by the Office of Naval Research.]

Acoustic vector sensors and vector/tensor processing in ocean acoustics research

Several passive sensor systems that simultaneously measure acoustic pressure and the three components of acoustic particle motion have been designed and deployed in seagoing experiments by the Marine Physical Lab over the past two decades. Descriptions of these sensor systems and results from applying various signal/array processing methods with these data are presented in this talk. Although vector sensors do not provide any additional information on the sound field than that obtained by a suitable number of properly placed hydrophones, they do provide estimates of the sound field's directional properties from a very compact package. Applications where available space is severely limited include autonomous underwater vehicles (AUV). Recent results from acoustic vector sensors installed in a flying wing underwater glider and inside a prop-driven AUV to quietly measure sounds in the ocean illustrate the benefits, and drawbacks, of the use of vector sensors on autonomous platforms. The extension of the signal/array processing approaches to tensor sensors indicates the promise of additional gain, if the sensor design challenges can be overcome. [Work supported by the Office of Naval Research].

Uni-vector-sensor direction finding with regularized ESPRIT

The regularized Least-Squares Estimation method of Signal Parameters via Rotational Invariance Techniques (LS-ESPRIT) is herein proposed for Direction-Of-Arrival (DOA) estimation of non-Gaussian sources with only one acoustic vector-sensor. The Second-Order Statistics (SOS) and Higher-Order Statistics (HOS) of data are fused within a regularization framework. The steering vectors can be blindly identified by the regularized ESPRIT, from which the aim of DOA estimation can be achieved. Several variants of the regularized ESPRIT are discussed. A suboptimal scheme for determination of the regularization parameters is also given.