ANI

Expanding uses of ambient noise for imaging, detection and communication: 

Expanding uses of ambient noise for imaging, detection and communication John R POTTER Acoustic Research Laboratory, Tropical Marine Science Institute, National University of Singapore & Laurent MALOD , EISM, France 3 December 2002

In This Presentation……: 

In This Presentation…… Some background to ANI - human history Acoustic Daylight™ and the first camera Expanding the algorithm base Expanding the hardware support A new tack; coherent ANI ideas Short-range, shrimp-powered? Longer-range, ship & whale-powered?

Background to ANI - Acoustic Daylight: 

Background to ANI - Acoustic Daylight Active sonar

Background to ANI - Acoustic Daylight: 

Background to ANI - Acoustic Daylight Active sonar Passive Sonar

Background to ANI - Acoustic Daylight: 

Background to ANI - Acoustic Daylight Active sonar Passive Sonar Acoustic Daylight, a novel concept in sonar. (Buckingham and Glegg)

Acoustic Daylight Ocean Noise Imaging System (ADONIS): 

Acoustic Daylight Ocean Noise Imaging System (ADONIS)

Ambient Noise Imaging: 

Ambient Noise Imaging ‘AD’ is but one statistical technique in a broad class of possible algorithms We might be able to use higher moments of the statistics Other algorithms may improve robustness and may not have any visual analogy

Example additional statistical algorithms: 

Example additional statistical algorithms AD - Intensity pdf mean Intensity pdf width Spatial Coherence imaging

Expanding the hardware support: 

Expanding the hardware support 1.4 m aperture 500+ broadband sensors 196 kSa/s per sensor Fully digital 1.6 Gbits/s dataflow 54 Pentium mProc data collectors Fibre-channel driven Fibre-optic data lines Not potted or oil-filled - helium partial pressure Remotely Operated Mobile Ambient Noise Imaging System

Not everything in life goes as planned….: 

Not everything in life goes as planned….

A new tack…. Coherent Ambient Noise imaging: 

A new tack…. Coherent Ambient Noise imaging What if we could opportunistically process sounds coherently instead of statistically? Obtain the random source positions from matched field estimations or similar processing. Then… Process these sources as for multi-static active sonar Radar precedent in ‘Silent Sentry’ Silent Sentry, passive surveillance radar - Lockheed Martin Mission Systems (1999) Y Wu & D. C Munson, University of Illinois-multistatic synthetic aperture radar imaging DSTO in Australia is building a coherent ANI array (Cato and Readhead)

Short-range: Shrimp-powered?: 

Short-range: Shrimp-powered?

Shrimp-powered active sonar…: 

Shrimp-powered active sonar… In shallow, warm water snapping shrimp dominate the spectrum from 2-300 kHz. Could we use snapping shrimp as illuminators for an ANI ‘multi-static active’ sonar?

Snapping Shrimp Noise-signal characteristics: 

Snapping Shrimp Noise-signal characteristics Bandwidth > 300 kHz, SL up to 180 dB re 1mPa2/Hz (Au, 1998) Reef fish are known to orient themselves to reefs by sound

The Concept Scenario: 

The Concept Scenario Snapping Shrimp as multi-static sonar illuminators Coherent processing of individual snaps Complicated by the lack of information about where and when the source occurred Target Receiver Sources

Distribution Characteristics: 

Distribution Characteristics Energy Density follows a log-normal distribution, a robust characteristic of snapping shrimp noise Time of arrival follows a Poisson Distribution (Miklovic, Tze wei)

Distribution characteristics- two spatial models: 

Distribution characteristics- two spatial models Distributed homogeneously & uncorrelated or Live in colony (spatially inhomogeneous) & chorus? Homogeneous distribution Colony

Limits on the processing range: 

Limits on the processing range Where can we treat signals coherently? Absorption & spreading losses will limit the useful range Larger range => snaps too close together and too low in amplitude

Scintillation index: 

Scintillation index S.I.= [<I2>-<I>2]/<I>2

Theoretical Formulation : 

Theoretical Formulation Unlike the conventional multi-static case, shrimp snaps are not well defined signals and have unknown location & time To formulate the imaging algorithm the following steps are required: Estimate the position & time of the signals Process data for potential ‘echoes’ Distinguish ‘echoes’ from other snaps

Source direction estimation: 

Source direction estimation Conventional beamforming of a planar array will provide source direction estimates for broadband sources For a large number of sensors (highly-resolving array) and high frequency (for high temporal resolution) conventional beamforming is computationally intensive. A technique based on phase contour pattern analysis is proposed to generate initial source direction estimates

Shrimp location from phase contour pattern-azimuth variation: 

Shrimp location from phase contour pattern-azimuth variation Shrimp along horizontal plane Shrimp along vertical plane Lines of equal phase Lines of equal phase

Phase contour pattern-Polar angle variation: 

Phase contour pattern-Polar angle variation Large polar angle case Small Polar angle case

Estimating  and : 

Estimating  and  Orientation of phase contours provides  estimate Spacing of phase contours provides  estimate Altitude of array provides range estimate Surface reflection provides geometric constraints to optimise estimates Spherical coordinates

Resolving the received signals: 

Resolving the received signals After the direct path signal is received, a new detected signal could be: A snap reflected off the bottom A snap reflected off the surface An echo reflected off a target … or A new snap A decision logic has to be developed to mimimise incorrect assignations

Resolving received signals (contd.): 

Resolving received signals (contd.) Physical and biological constraints help Echo must be smaller in amplitude than the original snap (and decreases with time) Surface reflections result in a phase shift and must come from a known upper half plane of the same azimuthal angle as the source. Since shrimp live on (or very near) the bottom, the bottom reflected signal is already included in the main arrival as an inseparable pair

Resolving the received signals (contd.): 

Resolving the received signals (contd.) A knowledge of temporal snap density gives an indication of the expected inter-snap interval. We can estimate the probability of any subsequent received signal being a new snap.

A snap example: 

A snap example Constraint line using estimated source range and spreading

Long-range: Ship & whale-powered?: 

Long-range: Ship & whale-powered?

Whale Anti-Collision System (WACS): 

Whale Anti-Collision System (WACS) The WACS concept is to instrument a corridor of safety for marine mammals, within which cetaceans can be detected, tracked & their positions notified to vessels using the corridor to permit timely course alterations. WACS consists of the following elements: A set of fixed, passive acoustic VLA buoys forming a two-dimensional spatial aperture. Buoy-to-shore communication system. Automated acoustic detection, and localisation. Land station-ship communication system. A ship data reception system.

Whale and ship-powered?: 

Whale and ship-powered? Vocalising whale - located passively

WACS processing: 

WACS processing Use the corridor of VLA’s as a passive sonar array to detect and track ships and vocalising marine mammals using: Matched field-like processing TRM-like individual source focussing Once sources are identified, use them to locate non-vocalising whales using: Coherent multi-static ANI

Acknowledgements: 

Acknowledgements We wish to recognise the many partners, both international and local, and the team of researchers at the ARL for the many inputs they have made to the ideas presented here. This work is supported by a DSTA contract to the Acoustic Research Laboratory, TMSI, NUS.

There’s more to me than meets the eye…: 

There’s more to me than meets the eye…