underwater counderwater coommunications · computational ocean acoustics april 19, 2007 underwater...

Underwater CoUnderwater Co

Ballardbjblair@April 19

WHOIE Adviser: James Preisig

April 19, 2007 Underwater CoBallard

ommunicationsommunications

d [email protected]

9, 2007

MIT Adviser: Art Baggeroer

ommunicationsd Blair

1

Back

BS in Electrical and CoBS in Electrical and CoCornell university 2002

MS in Electrical and CoJ h H ki 2005Johns Hopkins 2005

Hardware Engineer, JH

PhD Candidate, MIT/WApril 19, 2007 Underwater Co

Ballard

kground

omputer Engineeringomputer Engineering, 2

omputer Engineering,

HUAPL 2002-2005

WHOI Joint Programommunicationsd Blair

2

Introduction

O 0% f l• Ocean covers 70% of plan• 11,000 meters at deepest p


and Motivation

netpoint


3

Underwate

WHOApril 19, 2007 Underwater Co

Ballard

WHO

r Technology

OI 2006ommunicationsd Blair

4

OI, 2006

Comm

S t kSensor networks

Autonomous underwate

Gliders

Manned VehiclesManned Vehicles


unication

er vehicles (AUV)


5

Current A

ScienceScience− Geological / bathymetric su

Underwater archeology− Underwater archeology− Ocean current measuremen

Deep ocean exploration− Deep ocean explorationGovernment

Fish population manageme− Fish population manageme− Costal inspection

IndustryIndustry− Oil field discovery maintena


Applications

rveys

nt

ntnt

ance WHOI, 2005


6

Applications plann

Ocean observation systemOcean observation system− Costal observation

MilitaryMilitary− Submarine communications

Shi i ti− Ship inspectionNetworking

M bil t k (DA− Mobile sensor networks (DAVehicle deployment

− Multiple vehicles deployed s− Resource sharing among ve


ned / in development

mm

s (covert)

ARPA)ARPA)

simultaneouslyehicles


7

Example Comm


munication System

PLUSnet/Seaweb


8

Technology fo

RF (~1m range)RF ( 1m range)− Absorbed by seawater

Laser (~100m range)( g )− Hard to aim/control− High attenuation except for blue/g

Ult L F ( 100 kUltra Low Frequency (~100 km− Massive antennas (miles long)− Very narrowband (~50 Hz)y ( )− Not practical outside of navy

Cable− Expensive/hard to deploy mainta− Impractical for mobile work sites


r communication

green)m)

in


9

Acoustics is

Fairly low powerFairly low power− ~10-100W Tx − ~100 mW Rx100 mW Rx

Well studied− Cold war military funding

Compact− Small amount of hardware n

Current Best Solution


s the solution

WHOI Micromodem

needed


10

Acoustics

A ti iAcoustic wave is compthrough water medium


Background

i t liression wave traveling


11

Sound

S d f d 1500Speed of sound ~ 1500

Deep water profile

Schmidt, Computational Ocean Acoustics


d Profile

0 /0 m/s


12

Global Oc

Schmidt, Comput


cean Profile

tational Ocean Acoustics


13

Shallow w

Schmidt, Comp


water profile

putational Ocean Acoustics


14

Speed of Sou

V ti l d d fi• Vertical sound speed profi• the characteristics of the• the amount and importan• the amount of bottom int• the location and level of

• Horizontal Speed of SoundN li i i i h• Nonlinearities in channe


und Implications

l i tle impactse impulse responsence of surface scatteringteraction and lossshadow zones

d impactsll response


15

Propaga

Schmidt, C


ation Paths

Computational Ocean Acoustics


16

Mul

Mi lti th d tMicro-multipath due to Macro-multipath due top


tipath

h frough surfaceso environment

Sea surfaceSea surface

TxRx

seabed


17

Time vary

Time variation is due to:Time variation is due to:− Platform motion− Internal waves− Surface waves

Effects of time variabilityEffects of time variability− Doppler Shift

cuff cd =

− Time dilation/compression of the

Channel coherence times often

c

Channel coherence times oftenChannel quality can vary in < 1


ying channel

received signal

n << 1 secondn << 1 second. 1 second.


18

Acoustic FocusingTime-Varying Channel Impulse Response


Time (seconds)

g by Surface WavesDynamics of the first surface scattered arrival


19Preisig, 2006

Multipath and Time V

Ch l t ki dChannel tracking and q− Equalizer necessary an

Coding and interleavingCoding and interleaving

In networks, message r


Variability Implications

lit di ti i it lquality prediction is vitalnd beefy

gg

routing


20

Shado

Sometimes there is no direct path two points. All paths are either suno paths.pProblem with communications be

instruments in upwardly refractingwater, deep water).Problem with communications be

surface in a downwardly refractingwater and deep water).


w Zones

Clay and Medwin, ,“Acoustical Oceanography”

(unscattered) propagation between urface or bottom reflected or there are

etween two bottom mounted g environment (cold weather shallow

etween two points close to the g environment (warm weather shallow


21

Shadow Zone Exa


amples (Deep Water)


22Figures from J. Preisig

Ambie

• Ambient noise• Ambient noise – Passing ships, storms, breaking wav– p.s.d. decays as 20dB/decade ->N(f)

Primary natural sources− bubbles, rain, and biologic sources such

BubblesCan cause communications channel to di− Can cause communications channel to di

− Can increase surface scattering losses (u

− Attenuation in bubble cloud can be 20dB/

F d d t tt ti ( k 3− Freq dependent attenuation (peak near 3

− Can persist for minutes

− Cause sound (noise)


ent Noise

ves, seismic events )= 1010 f-2 Watts/Hz re 1μPa

as snapping shrimp

isappearisappear

up to 10dB per bounce)

/meter

30kH )30kHz)


23

No


oise

Figure from:

Stojanovic, j ,WUWNeT'06


24

Bubble Clou


ud Attenuation



Band

C t f t iCenter frequency typica

Typical Bandwidth ~ 5-

Channel is inherently by− Modulation essential for


dwidth

ll d 10 30kHally around 10-30kHz

15kHz

band-limitedr high rate communications


26

Path loss an

P th LPath Loss− Spherical Spreading ~ r− Cylindrical Spreading ~

Absorption ~ α(f)-r

Thorp’s formula (for sea− Thorp s formula (for sea

410044

111.0)(log10 2

2

2

2

++

+=

ff

fffα

41001 ++ ff


nd Absorption

r -1

r -0.5

a water):a water):

(dB/km) 003.0 000275.0 22 ++ f


27

Short Rang


e Attenuation



Long Rang

SNR( f ) =171+10log(P) − r α( f ) − 20log(h /2

60

70

80

90

1 km

30

40

50

SN

R (d

B) 10 km

50 km

-10

0

10

2050 km

100 km

0 5 10 15 20 25-20

f (kHz)

Figure courtesy of Costas


Figure courtesy of Costas

ge Bandwidth

⎛ ⎞2) −10log(r − h /2) −10log N( f )df

f −df / 2

f +df / 2

∫⎛

⎝ ⎜ ⎜

⎞

⎠ ⎟ ⎟ dB

Source: 20 Watts

h 500h = 500 m

30 35 40

s Pelekanakis


29

s Pelekanakis

Attenuation of S

Schmidt CApril 19, 2007 Underwater Co

Ballard

Schmidt, C

Sound in Seawater

Computational Ocean Acousticsommunicationsd Blair

30

Computational Ocean Acoustics

Bandwidth

M d l ti fModulation frequency m

System inherently wide

Frequency curtain effecq y− Form of covert commun

Might help with network− Might help with network


Implications

t b k t lmust be kept low

e-band

ctnicationsk routingk routing


31

Latency

P ti f dPropagation of sound s− Feedback might take se− Channel changing faste

Most underwater nodesC i ti T− Communications Tx pow

− Retransmissions costly


and Power

l th li htslower than lighteveral seconder than feedback

s battery powered( 10 100W)wer (~10-100W)


32

Example

Power A

WHOI Micromodem

DSP Tex10

Transmit Power

10 Typ

Receive 80 Power Wh

Data Rate 805 pFS

Daughter Card / Co-processor


FS

e Hardware

Amp

Micromodem in action

Micromodem Specifications xas Instruments TMS320C5416 0MHz low-power fixed point processor

Watts pical match to single omni-directional ceramic transducer.

milliwatts hile detecting or decoding an low rate FSK packet.

-5400 bps packet types supported. Data rates higher than 80bps SK i dditi l d t b i d


33

SK require additional co-processor card to be received.

Conc

C i ti i thCommunicating in the oce− Time varying channel− Inconsistent noise− Shadow zones− Bubbles− Latency

Many questions still left to − Communications is not well− Many opportunities for adva


clusions

i diffi ltan is difficult

be answered researched

ancesommunicationsd Blair

34

Acknowle

I ld lik t th k thI would like to thank thehelping me with this pre

− Jim Preisig− Costas Pelekanakis− Henrik Schmidt− Milica StojanovicMilica Stojanovic− WHOI Acoustic team


edgements

f ll i l fe following people for esentation:


35

My Re

C di f th dCoding for the underwa


esearch

t h later channel


36

underwater counderwater coommunications · computational ocean acoustics april 19, 2007 underwater...

Documents