Andrew J. Poulsen

Robust adaptive vector sensor processing in the presence of mismatch and finite sample support

We present analytical results which quantify the effect of system mismatch and finite sample support on acoustic vector sensor array performance. One noteworthy result is that the vector aspect of the array ldquodampensrdquo the effect of array mismatch, enabling deeper true nulls. This is accomplished because the variance of the vector sensor array spatial response (due to rotational, positional and filter gain/phase perturbations) decreases in the sidelobes, unlike arrays of omnidirectional hydrophones. When sensor orientation is measured within a reasonable tolerance, the beampattern variance dominates the average sidelobe power response. Our analysis also suggests that vector sensor array gain performance is less sensitive to rotational than to positional perturbations in the regions of interest. We analytically characterize the eigen-SNR threshold, which depends on the signal and noise covariance and the number of noise-only and signal-plus-noise snapshots, below which (asymptotically speaking) reliable detection using sample eigenvalue based techniques is not possible. Thus for a given number of snapshots, since the dimensionality of the snapshot in a vector sensor array is larger than that of a hydrophone-only array, the eigen-SNR detection threshold will be greater whenever the eigenvector information is discarded. We present processing techniques customized to the unique characteristics of vector sensors, which exploit information encoded in the sample eigenvectors and are robust to the mismatch and finite sample support issues. These methods include adaptive processing techniques with multiple white noise constraints.

Bearing Stabilization and Tracking for an AUV with an Acoustic Line Array

Passive underwater detection and tracking sonar systems using autonomous underwater vehicles (AUVs) have many important applications. Because of imperfections in vehicle control, it is common for an AUV to undergo significant yaw and pitch oscillations. As a result, it is important to compensate for the vehicle motion when generating true bearing estimates while using a rigidly attached acoustic line array. This paper describes full beam interpolation tracker and bearing stabilization algorithms that were implemented to address these issues on an intelligent AUV sonar sensor and tested during a subsequent sea trial with the goal of providing target bearing estimates to a target track estimation algorithm. These beam tracking and bearing stabilization algorithms can also be applied to the case of a flexible towed array with some additional modifications. Initial results indicate that this is an effective method of measuring stabilized true target bearings

Vector sensor array sensitivity and mismatch: Generalization of the Gilbert‐Morgan formula

The practical implementation of any sensing platform is susceptible to imperfections in system components. This mismatch or difference between the assumed and actual sensor configuration can significantly impact system performance. This paper addresses the sensitivity of an acoustic vector sensor array to system mismatch by generalizing the approach used by Gilbert and Morgan for an array with scalar, omnidirectional elements [E.N. Gilbert and S.P. Morgan, Bell Syst. Tech. J. 34, (1955)]. As such, the sensor orientation is not an issue because it does not affect performance for an array of omnidirectional sensors. Since vector sensors measure both the scalar acoustic pressure and acoustic particle velocity or acceleration, the sensor orientation must also be measured to place the vector measurement in a global reference frame. Here, theoretical expressions for the mean and variance of the vector sensor array spatial response are derived using a Gaussian perturbation model. Such analysis leads to insight i...

Acoustic noise properties in the rapidly changing Arctic Ocean

The Arctic Ocean is undergoing dramatic changes, the most apparent being the rapidly reducing extent and thickness of the summer ice cover. As has been well established over prior decades, the environmental acoustics of the ice-covered Arctic is dominated by two major effects: the highly inhomogeneous ice cover, and the monotonically upward refracting sound speed profile, the combination of which forces all sound paths to be exposed to strong scattering loss and the associated loss of coherence. In some portions of the Arctic Ocean, however, a persistent inflow of a shallow ‘tongue’ of warm Pacific water has recently been strengthening, which has dramatically altered the acoustic environment, creating a strong acoustic duct between approximately 100 and 200 m depth. This duct has the potential of trapping sound out to significant ranges (80-100 km) without interacting with the ice cover, resulting in much higher coherence and signal preservation. Acoustic noise measurement results collected with a vertica...

Changing Arctic ambient noise

The Arctic Ocean is undergoing dramatic changes, with the most apparent being the rapidly reducing extent and thickness of the summer ice cover. Furthermore, a persistent inflow of a shallow tongue of warm Pacific water has recently been discovered in the Beaufort Sea region of the Arctic, often called the Beaufort Lens, which creates a strong acoustic duct between approximately 100 and 200 m depth. These changes have had a significant effect on underwater acoustic propagation and noise properties. In spring 1994, acoustic data was collected in the Beaufort Sea region of the Arctic using a suspended vertical array; in spring 2016, similar data was collected in the same region. The 1994 data features meandering narrow-band features due to ice ridge friction, while the 2016 data in the new Arctic is largely dominated by ice mechanical events at discrete ranges and bearings. Supported by acoustic noise modeling, we illustrate these and other noise properties measured more than two decades apart in a region o...

Spatial diversity of ambient noise in the new Arctic

The Arctic Ocean is undergoing dramatic changes, the most apparent being the rapidly reducing extent and thickness of the summer ice cover. As has been well established over prior decades, the environmental acoustics of the ice-covered Arctic is dominated by two major effects: the highly inhomogeneous ice cover, and the monotonically upward refracting sound speed profile, the combination of which forces all sound paths to be exposed to strong scattering loss and the associated loss of coherence. In some portions of the Arctic Ocean, however, inflow of warm Pacific water has created the so-called “Beaufort Lens,” a neutrally buoyant high sound velocity layer at 70-80 meter depth, which has dramatically altered the acoustic environment, creating a strong acoustic duct between approximately 100 and 300 m depth. This duct has the potential of trapping sound out to significant ranges (80-100 km) without interacting with the ice cover, resulting in much higher coherence and signal preservation. Acoustic noise m...

Arctic ambient noise measurement comparisons in the Beaufort Sea

During ICEX 2016, an MIT autonomous underwater vehicle (AUV) was deployed in the Beaufort Sea region of the Arctic Ocean with the technical objectives of demonstrating the deployment, operation and recovery of an AUV with a towed array under extreme under-ice Arctic conditions, and the scientific objective of characterizing the acoustic environment. Part of this effort included suspending the AUV from a hydro-hole with the acoustic array hanging in a vertical configuration. This new data set created a choice opportunity to reprocess similar vertical array data from the same region of the Beaufort Sea collected by MIT during SIMI 1994. In the intervening twenty-two years between these two ice camps, significant changes have occurred in the Arctic, including rapidly reducing extent and thickness of the summer ice cover. Furthermore, a persistent inflow of a shallow “tongue” of warm Pacific water has recently been discovered in the Beaufort Sea, creating a strong acoustic duct between approximately 100 and 2...

Analysis of the advantages and complexities of acoustic vector sensor arrays

The hydrophone, an omnidirectional underwater microphone, is the most common sensor for listening to underwater sound. Directional sensors, however, have many important applications. Acoustic vector sensors, one important class of directional sensors, measure acoustic scalar pressure along with acoustic particle motion. With this additional vector measurement, vector sensors feature many advantages over conventional omnidirectional hydrophone sensors: improved array gain/detection performance, enhanced bearing resolution, the ability to “undersample” an acoustic wave without spatial aliasing, and the capability of attenuating spatial ambiguity lobes, e.g., left/right ambiguity resolution for a linear array. Along with their advantages, however, vector sensors also pose additional practical complexities: greater sensitivity to non-acoustic, motion-induced flow noise at low frequencies, requisite knowledge/measurement of each sensors orientation, management of different sensor types (pressure and particle motion) that each with different noise properties/calibration requirements, and adaptive processing can become difficult in a snapshot limited regime since each vector sensor is made up of up to four data channels. This paper will explore the virtues and limitations of vector sensor arrays in the presence of realistic ocean noise fields and system imperfections, including their effects on array performance (gain, beampatterns, etc.) supported by theoretical analysis and illustrative examples.The hydrophone, an omnidirectional underwater microphone, is the most common sensor for listening to underwater sound. Directional sensors, however, have many important applications. Acoustic vector sensors, one important class of directional sensors, measure acoustic scalar pressure along with acoustic particle motion. With this additional vector measurement, vector sensors feature many advantages over conventional omnidirectional hydrophone sensors: improved array gain/detection performance, enhanced bearing resolution, the ability to “undersample” an acoustic wave without spatial aliasing, and the capability of attenuating spatial ambiguity lobes, e.g., left/right ambiguity resolution for a linear array. Along with their advantages, however, vector sensors also pose additional practical complexities: greater sensitivity to non-acoustic, motion-induced flow noise at low frequencies, requisite knowledge/measurement of each sensors orientation, management of different sensor types (pressure and particle ...

Ambient noise modeling for high fidelity acoustic simulation

A high fidelity acoustic simulator has been developed to enable realistic simulation of array time series in order to test sensing/processing algorithms. This simulator, capable of generating calibrated hydrophone or vector sensor array data, is fully integrated into a three-dimensional hydrodynamic array model, generating sensor time series for dynamically evolving array shape, location, and orientation. Furthermore, the simulator handles a changing number and configuration of acoustic sources, targets, and receivers while utilizing the legacy ray tracing code BELLHOP to account for ocean multipath effects. Properly modeling ambient noise is particularly important when determining the effect of noise on sensor/processing performance, e.g., array gain can vary significantly based on the directionality of ambient noise. Embedded in the simulator is the capability to efficiently generate broadband noise for an arbitrary noise intensity distribution as a function of depth and elevation angle (azimuthally sym...

Array gain of vector sensors in ocean noise

By measuring acoustic particle motion, vector sensors provide important capabilities to ocean sensing systems, such as the elimination of left/right ambiguity for towed arrays, the ability to ‘‘undersample’’ an acoustic wave without spatial aliasing, and improved detection performance and array gain in ocean noise fields. Several different ocean noise models exist, including isotropic noise, directional noise (both point and spatially spread directional noise sources), and realistic surface noise in a stratified ocean environment [W. A. Kuperman and F. Ingenito, J. Acoust. Soc. Am. 67, 1988–1996 (1980)]. Theoretical expressions are derived for array data covariance matrices under different noise models, then used with optimal MVDR beamforming weights to analyze array gain. In order to better understand the advantages of vector sensor arrays, this paper also includes a performance comparison between vector and hydrophone arrays. Of further interest is array sensitivity and performance in the presence of sy...