Ocean Acoustic Laboratory at the

1
Ocean Acoustic Laboratory at the Pacific Missile
Range Facility (OAL at PMRF) February 27,
2003 Peter Stein James K. Lewis Jason
Rudzinsky Subramaniam Rajan Scientific
Solutions Incorporated Nashua, NH - Kalaheo,
HI www.scisol.com
2
Outline

• Project Objectives
• Unique aspects of the PMRF Range
• Work to Date
• Ocean Model
• Ocean-Acoustic Tomography Using Existing Range
  Assets
• Future Work

3
OBJECTIVE
To enable an Ocean Acoustic Laboratory (OAL) at
the Pacific Missile Range Facility (PMRF). In
this case we define an Ocean Acoustic Laboratory
as an area of ocean where the 4D ocean (time
variable range and depth dependent temperature
and salinity plus bathymetry) is well known in
near-real time. To provide widespread
dissemination of this range information
for System Modeling and Evaluation Training and
Debrief Ocean Awareness Research
4
Pacific Missile Range Facility

• BSURE
• Depths 2000 5000 m
• Area 22.5 x 40 nm
• 18 Receivers, 2 Sources
• BARSTUR
• Depths 600 2000 m
• Area 12 x 10 nm
• 42 Receivers, 6 Sources
• SWTR
• Depths 40 600 m
• Area 14 x 5 nm
• 118 Receivers, 10 Sources
• Source Bandwidth 8-11 kHz

Kauai
Hydrophones used for communications and
underwater tracking
5
OAL IMPLEMENTATION
6

• Work To Date
• Development and calibration of Hawaii tidal
  forcing model
• Development of high resolution (1-3 km) Kauai
  ocean circulation and surface wave models
• First Field Test (Aug 01) Assessed feasibility
  of tomography using range assets
• Demo of ocean model during RIMPAC (June02)
• Development and installation of 192 channel
  acoustic data acquisition system (ADAS) to be
  used in part to perpetually collect acoustic data
  to drive tomographic inversion
• Modification of existing acoustic propagation
  model for bottom-mounted sensors
• Demonstration of new optimized acoustic signaling
  scheme for improved tomographic inversion (Nov
  02)
• Partial derivation of data-assimilation scheme
  (Feb. 17-20, 2003)

7
Ocean Model
8
Acoustic Travel Time Tomography Using the
Existing Assets at PMRF

• Approach
• Use the bottom mounted sources and receivers to
  measure acoustic travel time fluctuations caused
  by significant oceanographic effects
• Combine the acoustic observations with the ocean
  model to produce a refined image of the
  regions sound-speed/thermal structure
• Complications
• Detected signals must be accurately assigned to
  particular ray-paths
• Travel times must be determined to roughly 1
  msec accuracy
• Changes in travel time due to surface waves are
  similar in magnitude to the changes caused by
  significant oceanographic processes
• What is the best way to assimilate the acoustic
  observations into the ocean model?
• How do we manage the measurement and model
  uncertainties?
• How do we ensure that our tomographic process is
  unbiased?

9
Simulated Acoustic Travel Time Variability
10
Inversion Approaches
Acoustic Measurements, y(t)
Data-Oriented - Ocean model is used to constrain
tomographic inversion
Tomographic Inversion XT(t) F-1 y(t), X-(t)
X-(t)
Ocean Model

• External Data
• Mesoscale Eddy
• Atmospherics
• Tides

XT(t)
Data Assimilation X(t) aX-(t) bXT(t)
t t Dt
X(t)
X-(t) Ocean Model State Before
Assimilation X(t) Ocean Model State After
Assimilation XT(t) Tomography Modeled Ocean
State y(t) Acoustic measurements
Model-Oriented - Acoustic measurements are used
to constrain ocean model
Acoustic Measurements, y(t)
Data Assimilation X(t) X-(t) Gy(t)
X(t)
X-(t)
Ocean Model

• External Data
• Mesoscale Eddy
• Atmospherics
• Tides

t t Dt
11
Acoustic Transects in Ocean Model Grid
12
Simulated Tomographic Inversion
Synthetic Ocean
Ocean Model
Sound Speed, C(R,z,t) t 0, 1, 120 hours
Acoustic Propagation Model
Acoustic Travel Times y(t)
C - CT
Random Noise Added to y(t)
Noisy Acoustic Travel Times yn(t)

• Average Sound Speed
• Point-to-point SS Variability Correlation
• Bathymetry

Tomographic Inversion
CT(R,z,t)
13
(No Transcript)
14
ADAS Description
Range Hydrophones

• Previous / Existing System
• Data written to digital tape
• 50 channels can be recorded simultaneously
• No transmit capabilities
• ADAS
• System Installed in Nov 02
• Software written in LABView
• Data written to .wav files
• All channels can be recorded simultaneously
• Arbitrary waveform generation capabilities
• Script based transmit/receive sequences
• Data reduction and signal processing software
  developed in Matlab
• Source control capabilities in near future
  (spring 03)
• When all 178 channels are recorded
  simultaneously, 35.6 Mb/sec written to disk

BSURE
Existing Acoustic Recorder/Reproducer (DSPCon,
Inc)
Fibre-Channel Hub
ADAS PC (SSI)
Source Control Box (SAIC Maripro)
Range Projectors
15
ADAS CONTROL PANEL
16

• Acoustic Signal Transmission / Processing
• Approach
• Transmit signal from source
• Apply a replica-correlation filter to receive
  data
• Detect, separate and assign ray arrivals
• Determine the travel time along each ray path
• Complications
• Need to filter out travel time changes due to
  surface waves and fluctuating internal
  inhomogeneities, therefore need to transmit as
  often as possible before the ocean state
  changes
• Cross-talk clipping - cant detect an arriving
  signal if transmitting at same time need to let
  signals clear out before transmitting

17
Transmit Scheme

• Want to transmit as often as possible
• Cant transmit when receive signal is expected

18

• November 2002
• ADAS Installed
• 300 Gb data collected to test system
• Improved signaling scheme tested
• Data Analysis Objectives
• Determine expected arrival time observation
  errors
• Determine if ocean noise (surface waves) can
  be filtered out of acoustic data
• Determine if acoustic models (required for
  inversion) are unbiased

19
Example of Short-Time Acoustic Variability

• 70 pings every 3.6 seconds
• 0.1 sec chirp (8-11 kHz)
• Shown are matched-filter outputs

20

• Measured Acoustic Travel Times

Ray 1
Ray 2 /- 0.5 msec
Ray 3 /- 1.5 msec
Ray 4 /- 4.0 msec

• Observations
• Pulse spreading increases with number of
  boundary interactions
• Doppler invariant signals show improved SNR

21
Tuning the Inversion Process
Ray 2
Ray 1

• Tuning Process

1. Compute rough tomo correction using only
surface path
3. Observe remaining bias in multipath travel
times
22
Moving Ahead

• 1. Implement the data assimilation process
  currently being derived (L. Vincent of ONR
  accepted planning letter, full proposal to come )
• 2. HFX Experiment
• Share ocean model output, acoustic measurements
  with HFX researchers
• Using suite of environmental data collected by
  various researchers
• Evaluate ocean model fidelity with and without
  assimilating acoustic observations
• Continue developing and refining tomographic
  process
• With UNH (Mayer, de Moustier, Kraft) funded
  Phase I SBIR
• Conduct high resolution (5m horizontal, 0.3 meter
  vertical) bathymetric survey on SWTR
• Determine benefit of ocean-acoustic tomography on
  multi-beam echo sounder data reduction, including
  sediment analysis
• Use high resolution bathymetry to improve forward
  model relating environment and acoustic travel
  time
• 3. Demonstrate utility of OAL products in a
  realistic scenarios (modeled or real) and present
  to Fleet, Industry, Acedemia

23
Summary of Products

• Development of High Resolution Ocean Model
• Predominantly automated
• Integration of wave model
• Products available on website
• Installation of digital data recorder at PMRF
• Improves data recording capability of range
• Data is instantly available for analysis
• Arbitrary transmission capability
• Innovative Research in Ocean-Acoustic Tomography
• Determined that tomography using bottom mounted
  acoustic transducers is feasible
• Inversions are being done on a very small scale
  in littoral
• Developed and validated an improved signal
  transmission scheme
• Completing derivation of a unique and innovative
  data assimilation scheme
• HFX Experiment