Indonesian J
ournal of Ele
c
trical Engin
eering and
Computer Sci
e
nce
Vol. 2, No. 2,
May 2016, pp
. 351 ~ 358
DOI: 10.115
9
1
/ijeecs.v2.i2.pp35
1-3
5
8
351
Re
cei
v
ed Fe
brua
ry 29, 20
16; Re
vised
April 25, 201
6; Acce
pted
May 1, 201
6
Experimental Multipath Delay Profile of Unde
rwater
Acoustic Communication Channel in Shallow Water
Yasin Yousif Al-Ab
oosi
1,2
Ahmad Zuri
Sha'ameri
1
1
F
a
cult
y
of Ele
c
trical Eng
i
ne
e
r
ing, Un
iversiti
T
e
knologi Ma
la
ysi
a
, 813
10, S
k
uda
i, Johor, Mala
ysi
a
2
F
a
cult
y
of Eng
i
ne
erin
g, Univ
e
r
sit
y
of Mustan
siri
yah, Ba
ghd
ad, Iraq
*Corres
p
o
ndi
n
g
author, e-ma
i
l
:
y
m
yasi
n2@
li
ve.utm.m
y
A
b
st
r
a
ct
Th
e
sh
al
l
o
w wa
te
r ch
an
ne
l
i
s
a
n
en
vi
ro
nm
en
t th
a
t
i
s
of p
a
r
t
icular
interest
to many res
ear
ch fiel
ds
.
An un
derw
a
te
r acoustic ch
ann
el is c
har
acteri
z
e
d
as
a multip
ath c
han
nel. T
i
me-
v
aryin
g
multi
p
ath
prop
agati
on is
one of the
ma
j
o
r factor
s that limit the acousti
c commu
n
ic
at
i
on perfor
m
anc
e in sha
l
l
o
w
w
a
ter.
T
h
is study
con
ducts tw
o un
de
rw
ater acoustic
exp
e
ri
me
nts in
T
anj
ung
Ba
lau
,
Johor, M
a
lays
ia. A tra
n
sduc
e
r
and a
hydro
p
hon
e are su
b
m
er
ge
d at
different de
pths
and se
par
ated
by
different
distanc
es. Lin
ear
freque
ncy mod
u
late
d (LF
M) pulses are ch
os
en as the main
transmit sig
nal
for the experi
m
e
n
ts. T
he cross-
correlation
bet
ween the trans
m
i
tted
and rec
e
ived signals r
epres
ents t
he im
puls
e response
of the channel
(mu
l
tipat
h pr
ofi
l
e). The r
e
su
lts show
that th
e
amplit
ude
of th
e succ
essive
p
a
ths w
ill
not r
a
pidly
d
e
cli
ne,
a
n
d
vice v
e
rsa, w
hen
the
dista
n
ce
betw
een
the s
end
er
an
d the
rece
iver
incre
a
ses. M
o
reov
er, the t
i
me
differenc
e bet
w
een the d
i
fferent paths w
i
l
l
be s
m
al
l in
th
e
case of dista
n
c
e incre
a
se. In
other w
o
rds, th
e
successiv
e
pat
hs w
ill conver
g
e
in time.
Ke
y
w
ords
: Un
derw
a
ter co
mmu
n
ic
ations,
multi-p
a
th, s
hall
o
w
w
a
ter chann
el, Ray mod
e
l, correlati
on
Copy
right
©
2016 In
stitu
t
e o
f
Ad
van
ced
En
g
i
n
eerin
g and
Scien
ce. All
rig
h
t
s reser
ve
d
.
1. Introduc
tion
Incre
a
si
ng int
e
re
st in the defen
se se
ct
or, off-sh
ore
oil indust
r
y,
and othe
r co
mmercial
operation
s
i
n
the un
de
rwate
r
e
n
vironment
m
a
kes u
nde
rwater research
more
pop
ular.
Electrom
agn
etic waves i
n
the und
erwat
e
r e
n
vi
ronm
e
n
t
are expo
se
d
to
high attenuation and can
only travel very sh
ort di
stances. Th
ere
f
ore, t
he only
way that na
vigation, com
m
unication, a
nd
other wi
rel
e
ss appli
c
atio
n
s
ca
n be do
ne is thr
oug
h aco
u
sti
c
m
e
thod
s [1-3].
The unde
rwater
aco
u
sti
c
com
m
unication
chann
el (UW-ACC) is di
fficult to employ
and
ha
s inh
e
rent
pro
b
le
ms.
The difficulty come
s
from
cha
nnel
characteri
stics,
such
a
s
atten
uation, m
u
ltipath fadin
g
, time-
varying
cha
r
a
c
teri
stics, an
d
cha
nnel i
nho
mogen
eiti
es [
4
]. The atten
uation of
sou
nd in the
oce
an
is a frequ
en
cy-de
pen
dent
pro
c
e
s
s. Hence, the o
c
ea
n a
c
ts a
s
a lo
w-pa
ss filter [5].
The
unde
rwater chann
el
sho
w
s in
homo
gen
eities i
n
sp
e
ed, tempe
r
at
ure,
and
sali
nity [1]. The
s
e
variable
s
m
a
y also
ch
ang
e in time
and
may be
diffe
rent fo
r the
same d
epth
of different
pla
c
es.
Therefore, th
e cha
nnel im
pulse re
spo
n
s
e chan
ge
s b
o
th spatially a
nd tempo
r
aril
y [4, 6].
Multipath o
c
curs i
n
UW-A
CC b
e
cau
s
e
of reflectio
n
s and refractio
n
s. Refle
c
tio
n
s o
c
cur
at the bottom
and th
e un
d
e
rwater
ch
an
nel surfa
c
e, whe
r
ea
s refra
c
tion
s
o
c
cur becau
se
of sound
cha
nnel
s
cre
a
ted by
the
sound
spee
d i
nhomo
gen
ei
ties. T
he
num
ber of m
u
ltipaths
re
aching
the
receiver si
de
can be very
large. Ho
we
ver, t
he multipath unde
r n
o
ise level is
ignored [2, 7].
Multipath sig
nals gen
eral
ly
r
epresent
aco
u
sti
c
e
nergy l
o
ss
.
Ho
weve
r, the inte
r-sym
bol
interference (ISI) will also
be det
ri
ment
al at the receiver in
comm
unication sy
stems because it
can
si
gnifican
t
ly increase t
he e
r
ror rate
of the
re
ceive
d
si
gnal.
The
time differen
c
e
between
the
last arrival p
a
t
h over the n
o
ise l
e
vel an
d the syn
c
h
r
o
n
ize
d
path i
n
the ch
ann
el that suffe
rs from
the multipath
is called d
e
la
y spre
ad [1]. The len
g
th of
this del
ay sp
read i
n
digital
comm
uni
cati
on
system
s with
out equali
z
ati
on pla
c
e
s
a l
o
we
r bo
und
on the du
rati
on of a symb
ol
or an u
p
p
e
r
boun
d on
the
data rate of t
he sy
stem th
at must b
e
u
s
ed to avoi
d chann
el-in
d
u
c
ed ISI [4, 8]. This
study aim
s
to determi
ne
the multipath
profile of
th
e UW-ACC i
n
shall
o
w
wa
ter at a different
rang
e.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
2, No. 2, May 2016 : 351 –
358
352
2. Under
w
a
t
er Cha
nnel Char
acteristi
cs
This
se
ction focu
se
s on
so
me paramete
r
s of t
he UW-ACC that affe
ct the aco
u
sti
c
sig
nal
prop
agatio
n from the tra
n
sducer to the h
y
droph
one.
2.1. Signal Atte
nua
tion
An aco
u
sti
c
signal
und
erwater exp
e
rien
ce
s att
enuatio
n du
e to sp
read
ing and
absorptio
n.
Path loss i
s
the me
asu
r
e of the l
o
st
sign
al inte
n
s
ity from the
proj
ecto
r to
the
hydrop
hon
e. Sprea
d
ing l
o
ss i
s
d
ue to t
he expa
ndin
g
are
a
that th
e so
und
sig
n
al en
comp
asse
s
as it geomet
ri
cally sp
rea
d
s
outwa
rd from
the sou
r
ce [9].
∗
1
0
l
o
g
(1)
whe
r
e
R
is th
e ran
ge in
m
e
ters
and
k
i
s
the sp
rea
d
in
g factor.
Wh
e
n
the medi
um
in whi
c
h
sign
al
transmissio
n occurs is un
b
ound
ed, the spreadi
ng is
sph
e
ri
cal and
the sprea
d
in
g factor
2
;
whe
r
ea
s in b
ound
ed sp
re
ading, it
is consi
dered a
s
cylindri
c
al
1
.
In prac
tic
e
, a s
p
reading
factor of
1
.
5
is
often c
o
ns
idered [2].
The ab
so
rptio
n
loss is a
re
pre
s
entatio
n
of
the energy loss i
n
form
of heat due t
o
visco
us
friction
an
d io
nic
relaxatio
n
that o
c
cur a
s
the
wave
g
enerated
by
an a
c
o
u
sti
c
signal
pro
pag
a
t
es
outwa
rd
s; this loss varie
s
l
i
nearly with
range a
s
follo
ws [9]:
,
1
0
l
o
g
∗
(2)
whe
r
e
is ran
ge in kilo
metres an
d
is the
abso
r
ption
coefficient. Th
e absorption
coeffici
ent
for frequ
enci
e
s above a f
e
w hun
dred Hz
can be e
x
presse
d em
pirically usin
g
Thorp’
s formula
[10], which d
e
fines
/
as a fu
nction of
.
0.
11
1
4
4
4100
2
.
7
5
.
1
0
0
.
0003
.1
0
(3)
Total path loss is the combi
ned contrib
u
tion of
both the
spre
adin
g
an
d absorptio
n losse
s
[1].
∗
10
log
1
0
l
o
g
∗
(4)
Since the pat
h losse
s
, expresse
d in dB, must be returned to its nat
ural value.
10
10
(5)
As seen i
n
Fi
gure
1, the p
a
th loss i
s
p
r
oportio
nal to
the ope
rating
frequ
en
cy. Figure
2
sho
w
s the
p
a
th lo
ss ve
rsus th
e rang
e
at diffe
re
nt
freque
nci
e
s,
the path l
o
ss increa
se
s
with
distan
ce
and
frequ
en
cy.The
shallo
w
water u
nde
rw
a
t
er
a
c
ou
stic cha
nnel ha
s highe
r
valu
es
of
attenuation t
han the
dee
p wate
r un
d
e
rwater
acou
stic
cha
nnel
[2]; while transmi
ssion l
o
ss
increa
se
s wit
h
distan
ce a
n
d
freque
ncy for both.
2.2. Sound Speed
The
spe
ed
of so
und
in
se
awater i
s
a
f
und
am
ental o
c
e
ano
grap
hic varia
b
le that
determi
ne
s th
e beh
avior of
sou
nd p
r
op
ag
ation in
the o
c
ea
n. Many e
m
piri
cal form
ulas
have be
en
develop
ed o
v
er the years for
cal
c
ula
t
ing so
und
speed u
s
in
g
values of
wa
ter tempe
r
at
ure,
salinity, and
pre
s
sure/dept
h. A simplifie
d expre
s
sion
for the so
und
spee
d was g
i
ven by Med
w
in
[3]:
1449
.2
4
.6
0
.
055
0
.0002
9
1.
34
0
.01
3
5
0
.016
(6)
whe
r
e
c
i
s
t
he
spee
d of
sou
nd in
sea
w
ater,
T
i
s
the
water te
mperature
(in
deg
ree
s
Cel
s
iu
s),
S
is the salinity (in
parts p
e
r tho
u
sa
nd) a
nd
d
is the depth (i
n meters).
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4
752
Expe
rim
ental Multipath Del
a
y Profile of
Und
e
rwate
r
Acou
stic Com
m
unication
…
(Yas
in Y.A.A.
)
353
Figure 1. Tra
n
smi
ssi
on lo
ss as a fun
c
tio
n
of
freque
ncy
Figure 2. Tra
n
smi
ssi
on lo
ss as a fun
c
tio
n
of
r
a
n
ge
3. Ra
y
Model
The ge
ometry of multipath pro
pagatio
n is
impo
rtan
t for commu
n
i
cation
syste
m
s which
use
array
pro
c
e
ssi
ng to
su
ppre
s
s m
u
ltip
ath. The
de
si
gn of
su
ch
sy
stem
s i
s
a
c
co
mpanie
d
by t
h
e
use
of a
p
r
o
pagatio
n mo
del fo
r p
r
edi
cting the
mult
i
path
config
uration. Acou
stic p
r
o
pagatio
n in
the o
c
ea
n i
s
govern
ed
by
the wave e
q
u
a
tion. As
sol
u
tions
to
the
wave equ
atio
n
a
r
e difficult
to
find in general cases, ap
proximatio
ns
are oft
en u
s
e
d
to model propa
gation [1
1, 12]. The ray
theory p
r
ovid
es o
ne
su
ch
approximatio
n. The s
hallo
w water
cha
n
nel was m
o
d
e
led a
s
a Pe
keri
s
waveg
u
ide, consi
s
ting
of an isovelo
c
ity layer ove
r
an
isovelo
c
ity ha
lf-spa
ce
as
shown in Fig
u
re
3.
Figure 3. Sch
e
matic
sho
w
i
ng a Peke
ri
s waveg
u
ide
m
odel of the sh
allow
water a
c
ou
stic
cha
n
n
e
l
whe
r
e
is the depth of
the
sou
r
ce,
is
the depth of th
e receive
r
,
is the height of the wate
r
colum
n
and
is the transmissi
on ra
ng
e. The distan
ce
travelle
d by the sound
along variou
s
eigen
rays ca
n be
comp
ute
d
u
s
ing
the
m
e
thod
of ima
g
e
s [7]. T
he
di
stan
ce
alon
g
dire
ct eig
e
n
r
ay
is den
oted by
given by:
(7)
whe
r
e
is the
distan
ce
alon
g an u
p
ward
origin
ating ei
genray with
as
su
rface ref
l
ection
s a
n
d
as bottom ref
l
ection
s. For
su
ch eig
enra
y
s
,
0
1
and:
2
1
(8)
whe
r
e
is the distance alo
ng a down
w
a
r
d origin
ating
eigenray with
as surfa
c
e reflectio
n
s
and
as bottom reflectio
n
s.
For su
ch ei
g
enray
s,
0
1
and:
2
1
(9)
1
2
3
4
5
6
7
8
9
10
40
41
42
43
44
45
46
47
F
r
equenc
y
V
S
.
T
r
ans
m
i
s
s
i
on
Lo
s
s
e
s
F
r
equenc
y
(
K
H
z
)
T
r
a
n
sm
i
s
si
o
n
L
o
ss
e
s
(
d
b
)
0
10
0
20
0
30
0
400
500
60
0
70
0
80
0
90
0
100
0
0
10
20
30
40
50
60
70
Ra
ng
e
V
S
.
T
r
a
n
s
m
i
s
s
i
on
L
o
s
s
e
s
Ra
ng
e(
m
)
Tr
a
n
s
m
i
s
s
i
o
n
Lo
s
s
es
(
d
b)
1K
hz
20
K
h
z
10
0K
hz
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
2, No. 2, May 2016 : 351 –
358
354
3.1. Reflec
ti
on at th
e Sea Surfac
e
The imp
edan
ce mi
smat
ch
betwe
en the
sea
wate
r an
d air
cau
s
e
s
t
he sea surfa
c
e to be
a very
goo
d reflector.
If the sea
su
rfa
c
e
is
cal
m
, t
he
reflectio
n
i
s
close
to p
e
rfe
c
t but in
clud
es a
pha
se shift by
π
radian
s i.e. the reflecti
on co
efficient
is
1
. If
the se
a su
rface is rough (due to
waves), a sm
all loss will be incurred
for every surface interaction [13].
e
x
p0.5
(10
)
whe
r
e
is the surfa
c
e
refle
c
tion coeffici
en
ts and
2
s
i
n
(11
)
0.324
.
1
0
.
(12
)
whe
r
e
is acoustic wave length (
2
,
is inci
dent angl
e,
is
rou
ghn
ess of the surfa
c
e
and
is the win
d
spe
ed in
/
.
3.2. Reflec
ti
on at th
e Sea Bottom
The im
ped
an
ce
mism
atch
betwe
en th
e
sea
water an
d seab
ed
ca
u
s
e
s
the
sea
b
o
ttom to
reflect
some
of the sou
nd i
n
cid
ents o
n
it. Where
and
are the d
e
n
s
i
t
y and sou
nd
spe
ed in
sea
water re
spe
c
t
i
vely
and
and
are the de
nsity and sou
nd sp
eed in t
he se
abe
d re
spe
c
tively.
For a sm
oot
h sea botto
m, the reflection is
angle
depend
ent and is given
by the Rayleigh
reflec
tion c
oeffic
i
ent [13, 14] as
:
cos
√
cos
√
(13
)
w
h
er
e
,
(14
)
The angle of
inciden
ce
can be co
m
puted ba
sed
on the geometry of the
Pekeri
s
waveg
u
ide. Let
angl
e
correspon
d to
an eig
enray
and a
ngle
co
rre
sp
ond
to an
eigen
ray
. Then, we have:
t
a
n
2
1
t
a
n
2
1
(15
)
3.3. Grazing
Angles
The a
ngle
wi
th whi
c
h
ea
ch ray
grazes the
b
oun
dari
e
s i
s
usually
termed
a
s
a
gra
z
ing
angle. Thi
s
is quite importa
nt becau
se of
its
influence
on both the bottom and su
rface
reflectio
n
c
oeffic
i
ents
[9].
∅9
0
(16
)
whe
r
e
is the incid
ent angl
e.
3.4. Propaga
tion Delay
The del
ay of each refle
c
te
d ray with respect to
the di
rect path i
s
rel
a
ted to differe
nce
s
in
the lengths of different paths. Let us cal
l
the propag
ation delay along the ray length
and
the propa
gati
on delay alo
n
g
the ray leng
th
.
,
(17
)
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4
752
Expe
rim
ental Multipath Del
a
y Profile of
Und
e
rwate
r
Acou
stic Com
m
unication
…
(Yas
in Y.A.A.
)
355
The propa
ga
tion delay of
se
cond
ary rays with
respect to the
dire
ct ray is a very
importa
nt parameter i
n
the
unde
rwater
cha
nnel
that affects syste
m
perfo
rma
n
c
e b
e
cause the
delayed
repli
c
a
s
of the receiver introdu
ce interf
e
r
en
ce
intersym
bol
hen
ce the
ne
ed to re
du
ce t
he
rate of transmis
s
i
on [4, 9].
3.5. Multipath Channel Model
Let
be the si
gnal tra
n
smi
s
sion th
rou
gh
the ch
annel
a
nd
be the co
rre
sp
ondi
ng
received si
gn
al. Ignoring t
he ab
solute
prop
agatio
n delay on the
dire
ct ray bet
wee
n
tran
smi
tter
and receiver
and combi
n
in
g the formul
a
s
, then
is expre
s
sed a
s
a
function of
in the
following way
[9]:
∞
∞
(18
)
The
ch
annel
impul
se
re
spo
n
se fo
r
a time-v
a
r
yin
g
multipath
unde
rwater aco
u
sti
c
cha
nnel
can
be expre
s
sed
as [2]:
,
(19
)
whe
r
e
and
denote time-varying path
amplitude and time-va
r
ying path delay
respe
c
tively.
Hen
c
e, ea
ch
path of an acousti
c ch
ann
el acts a
s
a lo
w-p
a
ss filter.
4. Chann
e
l Measur
e
men
t
s
Although it i
s
kno
w
n tha
t
the shallo
w water
ch
an
nel is d
o
min
a
ted by time
-varying
multipath, very few measu
r
ements of the
vari
ability of the multipat
h stru
cture
are done.
4.1. Sounding Signal
Since a unit impulse is an
unre
a
lizable
sign
al, engin
eers ch
oo
se a pra
c
tical in
put signal
to the system
that will lead to
an accurate estimation
of the sy
stem
impulse re
sponse. Several
sign
als
are
freque
ntly emp
l
oyed,
those
being
LFM
(li
near freq
uen
cy mod
u
lated
)
chirp
[15], white
noise [16], and DSSS BPSK (direc
t
s
equenc
e
s
p
read
s
p
ec
trum binary phase
s
h
ift k
e
ying) s
i
gnal
[15, 17]. All
si
gnal
s p
o
sse
s
s a
c
cepta
b
le
autocorrela
tio
n
prope
rtie
s
as to
ap
proxi
m
ate Di
ra
c
d
e
lta
function
clo
s
ely. Autocorrelation that
approximat
e
s
the Di
ra
c d
e
lta functio
n
is the te
st
o
f
good
ne
ss fo
r a so
undi
ng
sign
al. LFM
chirp si
gnal i
s
le
ss li
kely
to appe
ar
ra
ndomly in a
n
y
environ
ment
than the oth
e
r sig
nal
s [18]. Ther
efore
,
the LFM chirp si
gnal
was u
s
ed a
s
the
sou
ndin
g
sig
nal in the exp
e
rime
nts presented later in
this pap
er.
4.2. Impulse Resp
onse
For a
n
y line
a
r
sy
stem
with
impul
se
re
sp
onse
and in
p
u
t
, the outp
u
t
ca
n
be
found by usi
n
g convol
ution
betwee
n
inp
u
t
and impul
se respon
se of
system [19].
∗
(20)
The i
nput-ou
t
put rel
a
tionship d
e
fined
according
to
the
co
rrel
a
tion fun
c
tion
and th
e
power spe
c
trum is [20]:
∗
∗
.
.
|
|
(21)
∗
.
.
(22)
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
2, No. 2, May 2016 : 351 –
358
356
All the above
relation
ship
s apply to sa
mple fun
c
tion
s and di
screte-time si
gnal
s. Whe
n
the input sig
nal is LFM si
gnal, the autoco
r
relati
on
of this signal
is unit impulse si
gnal an
d
the
Fouri
e
r tra
n
sf
orm (F.T.
)
of unit impulse
signal i
s
eq
ual to one. T
herefo
r
e, the
result of cro
s
s-
correl
ation b
e
t
ween
output
and in
put re
pre
s
ent th
e i
m
pulse respo
n
se
of the
ch
annel
(multip
a
th
delay profile
).
5. Sea Trial
5.1. Experimental Setup
The ch
ann
el measurement
s we
re co
nd
uct
ed on the
18th of June, 2014 (10
a.m.–12
p.m.) in Ta
nj
ung Bala
u, Johor, M
a
laysi
a
(latitude:
1°
33.169
′
N; lo
ngitude: 1
04°
26.027
′
E). T
he
cho
s
e
n
lo
cati
on h
a
s a
n
av
erag
e
depth
of abo
ut 2
2
m. Tra
n
smissions were m
a
de from
an
o
m
ni-
dire
ctional
transdu
cer BII-803
0 u
nde
rwater a
c
ou
st
ic tran
smitter with
an
ava
ilable frequ
e
n
cy
rang
e of
20
Hz to 1
00
kHz
and a
maxim
u
m cable
len
g
th of 10
m.
The
sign
al was
re
ceived
u
s
ing
a bro
adba
nd
hydrop
hon
e (7 Hz–22
kHz) model
Dolp
hinEAR 10
0 Serie
s
). The
maximum ca
ble
length
wa
s 3
0
m Figu
re 4
.
The win
d
speed
wa
s
ab
out 7 knots.
The temp
erat
ure at the
se
a
surfa
c
e
wa
s 28
°C, while the salinity wa
s 35 ppt.
Similarly, the spe
ed of soun
d wa
s 1541.3
m/s,
as obtai
ned u
s
ing the Me
d
w
in eq
uation.
The tra
n
smitt
ed sig
nal u
s
e
d
wa
s a 3
0
m
s
linea
r fre
q
u
ency mo
dulat
ion (L
FM)
sig
nal with
a band
width
of 20 kHz cente
r
ed a
r
oun
d 40
kHz. Th
e different tra
n
sm
issi
on lo
cati
ons
corre
s
p
ond
ed
to 10
m an
d
100
m rang
es. Th
e
re
cei
v
ed si
gnal
was
sam
p
led
at 44
kbit/s
a
n
d
store
d
for late
r analysi
s
. Fi
gure 5
sho
w
s the experime
n
t site.
Figure 4. Con
f
iguration of the experi
m
en
t
at Tanjung B
a
lau, Jo
hor,
Malaysia o
n
June
18, 2014
Figure 5. Experime
n
t test site.
5.2. Results
The first te
st wa
s cond
uct
ed with
a 10
m ran
ge at 2
2
m depth. T
he tran
sd
uce
r
and th
e
hydrop
hon
e
were subme
r
ged at 5 m d
epth. The LF
M wa
s se
nt, afterwhi
ch, a
nd the re
ceiv
ed
sign
al wa
s re
corded at the
receive
r
sid
e
.
The
cro
s
s-correlation p
r
o
c
e
ss was p
e
rformed bet
we
en
the tran
smitted and
re
cei
v
ed sign
als u
s
ing MAT
L
AB to obtain the path d
e
la
y profile. Fig
u
re 6
sho
w
s the multipath profil
e for 10 m ra
nge. The fi
gu
re cle
a
rly sh
o
w
ed that the gra
z
ing a
ngle
will
increa
se
wh
e
n
the di
stan
ce between
th
e se
nde
r
a
n
d
the re
ceive
r
sho
r
tene
d, th
ereby
de
cre
a
s
ing
the reflectio
n
co
efficient.
This re
sult
l
ed the
am
pli
t
ude of
the
su
ccessive
p
a
ths to
q
u
ickly
decrea
s
e a
n
d
disap
pea
r.
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4
752
Expe
rim
ental Multipath Del
a
y Profile of
Und
e
rwate
r
Acou
stic Com
m
unication
…
(Yas
in Y.A.A.
)
357
Figure 6. Multipath inten
s
ity profile at 10 m rang
e
Figure 7 sh
o
w
s the
re
sult
s of the se
co
nd
test with
100 m ra
nge
and 22 m d
epth. The
transdu
ce
r and the hydro
phon
e we
re subm
erged a
t
10 m depth. The figure sho
w
e
d
that the
numbe
r of
p
a
ths
and th
e
dista
n
ce bet
wee
n
the
se
nder an
d the
re
ceive
r
in
creased
whe
n
the
rang
e in
crea
sed. M
o
re
over, the g
r
a
z
in
g angl
e w
ill
decrea
s
e, th
ereby in
crea
sing th
e refle
c
tion
coeffici
ent. T
h
is result imp
lied that the a
m
plitude
of th
e su
cce
ssive
paths
will n
o
t rapidly d
e
cli
n
e.
The time
diff
eren
ce
b
e
tween th
e
different path
s
w
ill
be
sm
all. In
othe
r
wo
rd
s, the
su
cce
ssi
ve
paths
will converge in time.
Figure 7. Multipath inten
s
ity profile at 100
m range.
Table
1
sh
ows a
comp
ari
s
on of
the
sim
u
lated
delay
depe
nding
o
n
Eqs.
(7
), (8),
(9
), a
nd
(17
)
a
gain
s
t
the expe
rime
ntal propa
gat
ion d
e
lay
an
d the
amplitu
de at
ea
ch
p
a
th arrival. T
he
delay of e
a
ch
refle
c
ted
ray
wa
s d
e
termi
ned
with resp
ect to the
direct path.
The
observatio
n
for
the 100 m ra
nge wa
s ap
p
r
oximately the same a
s
we
ll as the pro
pagatio
n dela
y
for rays hitting
the surfa
c
e
o
r
b
o
ttom, su
rf
ace
–bottom–
surfa
c
e,
or b
o
ttom–su
rfa
c
e–bottom,
a
m
ong
othe
rs.
Thi
s
result wa
s ca
use
d
by the locatio
n
of both the
tran
smit
ter and the
re
ceiver
being
at nearly half
of
the cha
nnel d
epth.
Table 1. Co
m
pari
s
on of the
simulated a
g
a
inst expe
rim
ental pro
pag
a
t
ion delay
Ar
r
i
ve
num
ber
Range
(10
m)
Range
(100
m)
Experime
nt
dela
y
(
)
Simula
tion
dela
y
(
)
A
m
pl
it
ud
e
Experime
nt
dela
y
(
)
Simula
tion
Dela
y
(
A
m
pl
it
ud
e
1
0 0
1 0 0
0.7
2
3.75
2.97
0.467
3.125
2.28
1
3 15.037
15.212
0.1
7
6.1
0.7517
4
10.375
12.2
0.7122
5
11.625
12.6
0.805
6
16.375
19.3
0.3018
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
2, No. 2, May 2016 : 351 –
358
358
6. Conclusio
n
The study consi
ders
th
e
sh
ort-ran
g
e
sh
allow wa
ter hyd
r
oa
co
ustic chan
ne
l. The
multipath effect in a shall
o
w wate
r ch
an
nel ha
s also been di
scu
s
sed. The num
erical expe
ri
ment
results
clearl
y show
spati
a
l variability in the acou
sti
c
signal
s, whi
c
h are
required for
the design
of shallo
w
water
com
m
unication sy
stem
s. Mu
ltipath propa
g
a
tion ke
ep
s many hurdl
es in
achi
eving hig
h
data
rates and ro
bu
st
communi
cati
o
n
lin
ks.
The
delay
sp
read
of the
chan
nel
signifi
cantly decrea
s
e
s
in the ca
se of short-dista
n
ce
links. More
o
v
er, the effective data rate of
this ch
ann
el increa
se
s at a few hund
red
symbol
s pe
r se
con
d
.
Ackn
o
w
l
e
dg
ements
The auth
o
rs
woul
d like to t
han
k the Uni
v
ersiti Te
knol
ogi Malay
s
ia
(UT
M
) a
nd M
i
nistry of
High
er Edu
c
a
t
ion (MO
H
E)
Malaysia fo
r sup
portin
g
this wo
rk.
Referen
ces
[1]
G Burro
w
e
s
an
d JY Kh
an.
Sh
ort-rang
e u
nde
rw
ater acous
tic
co
mmun
icatio
n netw
o
rks
. INT
E
CH Open
Access Publ
ish
e
r. 2011.
[2]
T
Melodia,
H
Kulh
and
ji
an, L
C
Ku
o a
nd E
Demirors.
"Adv
ances
in
un
der
w
a
t
e
r
acoustic
net
w
o
rkin
g"
.
Mobil
e
Ad Hoc
Netw
orking: C
u
tting Edg
e
Dir
ections
. 20
13:
804-
852.
[3]
H Med
w
i
n
an
d CS Cla
y.
F
und
amenta
l
s of acoustica
l oce
a
n
ogra
phy
. Acad
emic Press. 19
97.
[4]
M Stojan
ovic
a
nd J Pre
i
si
g. "
U
nd
er
w
a
ter
ac
ous
tic comm
un
icatio
n ch
ann
el
s: Propag
atio
n
mode
ls an
d
statistical char
acterizati
on".
Communic
a
tions
Maga
z
i
ne, IEEE.
2009; 47: 84
-89.
[5]
M Chitre, JR
Potter an
d SH
Ong. "Optima
l
an
d n
ear-
opti
m
al si
gna
l d
e
tection
in s
n
a
p
p
in
g shrim
p
domi
nated ambie
n
t
nois
e
".
IEEE Journa
l of Oceanic En
gi
n
eeri
n
g
.
20
06; 3
1
: 497-5
03.
[6]
M Cale
y a
nd A
Dunca
n
. "Inve
s
tigatio
n of un
d
e
r
w
ater ac
ousti
c multi-path D
o
ppl
er an
d del
a
y
spr
ead
in
g
in a sha
llo
w
m
a
rin
e
envir
onm
ent".
Acoustics Australia.
20
1
3
; 41: 20-2
8
.
[7]
Y F
e
i, L Xiao-
Yang, W
Qian and C En. "Un
der
w
a
t
e
r Acou
stic Communic
a
tion Bas
ed o
n
H
y
perb
o
l
i
c
F
r
eque
nc
y Mo
dul
ated M-
ar
y Binar
y Orth
o
gon
al K
e
yin
g
".
T
E
LKOMNIKA Indo
nesi
a
n
Journ
a
l
of
Electrical E
ngi
neer
ing.
2
014; 12:
731
1-7
317.
[8]
B Boro
w
ski. "
Char
acteri
z
a
ti
o
n
of a very
shall
o
w
w
a
ter acoustic co
mmu
n
icati
on
chan
nel
". in
Procee
din
g
s of
MT
S/IEEE OC
EANS. 2009.
[9]
F
De R
a
n
go, F
Veltri
an
d P F
a
zio. "A
multi
p
ath fad
i
n
g
ch
a
nne
l mo
del
for
und
er
w
a
ter
sh
allo
w
a
c
ousti
c
communic
a
tio
n
s
". in
IEEE International Conference
on Comm
unic
ations (I
CC).
201
2: 381
1-38
15.
[10]
W
H
T
horp.
"A
nal
ytic
D
e
scri
p
tion of the
Lo
w
-
F
r
e
q
u
enc
y Attenuatio
n
C
o
efficient".
T
he
Journ
a
l of the
Acoustica
l Soci
ety of Americ
a.
1967; 4
2
: 270-
270.
[11]
F
B
Jensen.
Co
mp
utatio
nal oc
ean
ac
oustics
.
Sprin
ger Scie
n
c
e & Business
Medi
a. 199
4.
[12]
PC Etter.
Unde
rw
ater acoustic
mod
e
l
i
ng a
nd
simulati
on
. CR
C Press. 2013.
[13]
RP
Hod
ges.
U
nderw
a
ter ac
o
u
stics: Analysi
s, design a
nd
perfor
m
a
n
ce of
sonar
. John
W
ile
y
& S
ons.
201
1.
[14]
LM Brek
hovski
kh a
n
d
IUrP
L
ysanov.
F
u
nd
a
m
e
n
tals
of oce
an aco
u
stics
. S
p
rin
ger Sc
ie
nc
e & B
u
si
ness
Medi
a. 200
3.
[15]
S Dessa
lermo
s. "Und
ersea
acoustic
pro
p
a
gatio
n
ch
an
nel
estimatio
n
".
Montere
y
, C
a
li
fornia. N
a
va
l
Postgrad
uate Schoo
l.
200
5.
[16]
P Schomer. "Measur
ement
of Sound T
r
ansmissio
n
Los
s b
y
Com
b
in
i
ng Corre
lati
on
and F
ourie
r
T
e
chniques
".
T
he Jour
nal
of the Acoustic
a
l
Society of Ame
r
ica
.
197
2; 51: 112
7-11
41.
[17]
M Chitre, J Po
tter and OS H
eng. "U
nd
er
w
a
ter
acoustic c
h
ann
el ch
aracte
risatio
n
for me
dium-ra
ng
e
shall
o
w
w
a
t
e
r communic
a
tio
n
s
". in
OCEANS'04. MTTS/IEEE TECHNO-OCEAN'04
. 20
0
4
: 40-45.
[18]
Y F
e
i, W
W
en-Jun a
nd
C E
n
. "Ch
a
racteris
ti
cs Anal
ys
is o
f
HF
M Signa
l
over U
nder
w
a
t
e
r Acousti
c
Cha
nne
ls".
T
E
LKOMNIKA Indon
esia
n Jour
nal
of Electric
al
Engin
eeri
n
g
.
2
013; 11: 1
173-
118
0.
[19]
AV Oppen
heim
and GC Verg
h
e
se. "S
ig
nals,
s
y
stems, an
d i
n
ferenc
e".
Clas
s
notes for.
2010; 6.
[20]
PZ
Peebl
es, J
Rea
d
a
nd P
Rea
d
.
Prob
a
b
ility, ra
nd
om
varia
b
les,
and
rand
o
m
si
gn
al pr
inci
ples
.
McGra
w
-
Hil
l N
e
w
Y
o
rk. 200
1;
3.
Evaluation Warning : The document was created with Spire.PDF for Python.