Acoustic Communications in Shallow Waters: Sławomir Jastrzębski

ACOUSTIC COMMUNICATIONS IN SHALLOW WATERS
SŁAWOMIR JASTRZĘBSKI
University of Technology and Agriculture in Bydgoszcz

7 Kalisky Ave, 85-792 Bydgoszcz, Poland
email: [email protected]
The shallow water channel is an environment which is of particular interest to many

research workers. An underwater acoustic channel is characterized as a multipath channel.
Severe signal degradation can occur in such a channel due to multipath effects and the
refractive properties of the channel, which may include multiple interactions with the sea
bottom and sea surface. Time-varying multipath propagation is one of the major factors that
limit acoustic communication performance in shallow water.
In this paper the results of an analysis of broadband shallow water acoustic signal was
shown. Linear FM pulses were chosen as the main transmit signal for numerical experiments.
The transmitted signal goes through multiple paths in order to reach the receiver. The
numerical simulations were performed using the broadband normal mode model. This
software channel simulator is capable of simulating the effect of the shallow water
propagation channel on an input broadband acoustic signal and can be used for simulations
of sound propagation through both a time- and range-varing propagation channel. The
received signals were all matched filtered revealing the multipath arrival structure.
INTRODUCTION
Acoustic communications in shallow water had been a difficult problem due to the
channel characteristics of the underwater acoustic channel. For long range underwater
acoustic communications the main problem encountered is the presence of multipath
propagation caused by reflection and scattering of the transmitted signals at the bottom and
the surface.
Reflections from channel boundaries and diverse objects dominate the multipath structure.
The transmitted signal can go through multiple paths in order to reach the receiver. These
multiple paths can cause significant time spread in received signal. Each path has can possibly
have multiple surface interactions causing additional frequency spreading due to motion of
the water. Shallow water propagation is very sensitive to changes in the geometrical
parameters like water depth, source-receiver range or bottom slope leading to variations in the
impulse response of the underwater acoustic sound channel. Normal mode approaches have
been widely used in underwater acoustics and are derived from an integral representation of
the wave equation. When propagation is described in terms of normal modes, changed in the
environment translate into energy transfer between modes. In this paper the numerical
simulations were performed using the Prosim broadband normal mode model working in
Matlab environment [8].
1. BROADBAND NORMAL MODE MODEL

The Prosim can be used for simulations of broadband sound propagation through both a
time- and range-varying propagation channel. The broadband model is based on a model
called ORCA. ORCA is based on a layered normal mode approach assuming that the inverse
of the sound speed squared varies linearly with depth in each layer. This software project is
mainly intended for shallow water applications. The software consists of oceanographic and
acoustic models which perform specific tasks to describe the complexity of the real ocean.
The combination of several models can be a unique tool for predicting underwater acoustic
systems performances, such as communication systems for which channel stability and fading
are major problems. In this paper the acoustic part of package was used. This package
consists of a broadband normal mode code for calculating the received signals in the time
domain and this signals can be matched filtered in the case of FM transmissions.
The normal mode solution involves solving a one-dimensional equation. Normal modes
are normally thought of principally in the context of range-independent problems, but they
can be extended in various ways to both range-dependent problems and fully three-
dimensional problems. In this paper range-independent problem is considered. Figure 1 shows
the model of the range-independent environment [6].
r
zs
source c(z)
ρ(z)
Fig.1 Model of the range-independent environment

The acoustic wave equation has the following form:
⎛1 ⎞ 1 δ ( z − z s )δ (r )
∇ ⋅ ⎜⎜ ∇P ⎟⎟ − 2 Ptt = − s (t ) (1)
⎝ρ ⎠ ρc ( z ) 2πr
where P(r,z,t) is the acoustic pressure, s(t) is the point source, ρ(z) is the density and c(z) is
the sound speed.
Assuming that source time series has the form:
s (t ) = e − iωt (2)
then P(r,z,t) has the form:
P (r , z , t ) = p (r , z )e − iωt (3)
and we can write the following Helmholtz equation:
1 ∂ ⎛ ∂p ⎞ ∂ ⎛ 1 ∂p ⎞ ω 2 − δ ( z − z s )δ (r )
⎜ r ⎟ + ρ ( z ) ⎜⎜ ⎟⎟ + 2 p= (4)
r ∂r ⎝ ∂r ⎠ ∂z ⎝ ρ ( z ) ∂z ⎠ c ( z ) 2πr
After separating variables, i.e. substitution p(r,z)=Z(z)R(r) (with the source removed)
equation (4) becomes:
⎡ 1 ∂ ⎛ ∂R ⎞⎤ ⎡ ∂ ⎛ 1 ∂Z ⎞ ω 2 ⎤
+
⎢ r ∂r ⎜ ∂r ⎟⎥ ⎢
r ρ ( z ) ⎜ ⎟+ p⎥ = 0 (5)
⎣ ⎝ ⎠⎦ ⎣ ∂z ⎜⎝ ρ ( z ) ∂z ⎟⎠ c 2 ( z ) ⎦
Equation (5) leads to the modal equation after denoting this separation constant by k2:
d ⎛ 1 dZ ( z ) ⎞ ⎛ ω 2 ⎞
ρ ( z) ⎜⎜ ⎟⎟ + ⎜⎜ 2 − k 2 ⎟⎟ Z ( z ) = 0 (6)
dz ⎝ ρ ( z ) dz ⎠ ⎝ c ( z ) ⎠
Z (0) = 0
dZ
( D) = 0
dz
The modal equation has an infinite number of solutions which are characterized by a mode
shape function Zm(z) and a horizontal propagation constant km. The function Zm(z) is an
eigenfunction and km or k2m is an eigenvalue. Finally the pressure can be writing as a sum of
the normal modes:
− iπ ∞
i eikmr
p(r , z ) ≈ e ∑ Z m ( zs )Z m ( z)
4
(7)
ρ ( z s ) 8πr m =1 km
The Prosim propagation model is able to approximate a source beam with a Gaussian
shape, constant for all the frequencies, decomposing it as a sum of the normal modes each
multiplied by the appropriate shading coefficient. The beam is parameterized by the aperture
and tilt. The Prosim broadband normal mode model is in more detail described in [5,6].
2. PULSE COMPRESSION TECHNIQUES

Chirp pluses were used in these numerical experiments. Chirp is the nickname of Linear
Frequency Modulation (LFM). It is a wideband and long duration frequency sweep. It is a
noise-like signal. The differences between a chirp and a white noise signal are primarily the
phase changes in chirp is linear, the phase change of a noise signal is random. A frequency
band of chirp signal is limited whereas a noise signal’s frequency band is unlimited (“white”).
The chirp signal can be write in the following form in the time domain:
⎧ ⎛ B 2⎞ T
⎪⎪ A0 cos 2π ⎜ f 0t + 2T t ⎟ t ≤ 2
⎝ ⎠ (8)
⎨
⎪0 T
⎪⎩ t >
2
where T is the time duration of signal, f0 is the center frequency and B is the bandwidth of
signal. Spectrum of the signal after compression is given by:
Y ( f ) = H N ( f ) ⋅ HO ( f ) (9)
where Y(f) is the spectrum of the signal after compression, HN(f) is the spectrum of the
received signal and HO is the spectrum of the compression system. Signal y(t) in the time
domain has the form:
∞
y (t ) = ∫H
−∞
N ( f ) H O ( f )e jωt df (10)
In case of matched filtering:

H O ( f ) = H N∗ ( f )
thus: (11)
∞
sin(πBf 0t ) j 2πf0t
y (t ) =
−∞
∫H N ( f ) H N∗ ( f )e j 2πft df ≈
πBf 0 t
e (12)
Through matched filtering of the received signal with a transmit replica the pulse compression
is obtained.
3. NUMERICAL EXPERIMENT
In these numerical experiments a Hanning weighted chirps of bandwidth 2 kHz and
center frequency of 6 kHz with pulse duration of 500 ms was transmitted. The maximum
source level was 200 dB ref. 1μPa at 1m (Fig.2).
Fig.2 Transmitted chirp pulse in the time and frequency domain

Pulse compression was achieved through the use of the matched filter. The received
signals were correlated with a chirp replica of the transmit waveform. The matched filtered
data was used in order to examine the arrival time structure of the multipath.
The numerical experiments were repeated at ranges of 3, 6 and 9 km and with source
depth of either 30 m. The receivers were set at water depths of 30, 50, 70 and 90 m
respectively. Figure 3 shows the model of tested underwater acoustic channel and figure 4
shows sound speed profile for this channel. The range-independent calculations were
performed.
r = 3, 6 and 9 km
SEA SURFACE
hr = 30m
hs = 30m 50m
WATER
Tx 70m
90m
h1 = 100m α1=0.0dB/λ
ρ1=1.0g/cm3 Rx
c1 (table1)
α2=0.25dB/λ
h2 = 30m SEDIMENT ρ2=1.2g/cm3
SUBBOTTOM α3=0.5dB/λ ρ3=2.0g/cm3 c3=1800m/s
Fig.3 The model of shallow water hydroacoustic channel

The area of model is 100m in depth. The normal mode model assumes the bottom to consist
of single sediment overlying a homogenous sub-bottom. The thickness of sediment layer is
30m. The sediment parameters were set: sound speed linearly increasing from 1580 to
1620m/s, attenuation constant 0.25dB/λ and relative density 1.2 (relative to 1g/cm3, i.e.
relative to the water value). In numerical experiment there is a homogenous sub-bottom with
sound speed, relative density and attenuation of 1800m/s, 2.0 and 0.5dB/λ, respectively. Table
1 shows the value of sound speed profile for tested shallow water hydroacoustic channel.
Fig.4 Sound speed profile for tested underwater acoustic channel
Tab.1 The shallow water sound speed profile
Depth [m] Sound speed [m/s]

0 1522.1
8 1522.1
15 1514
69 1509.6
100 1509.6
Figures 5, 6 and 7 give examples of received signals at a range of 3, 6 and 9 km and
depths of 30, 50, 70 and 90m, respectively. Figures 8, 9 and 10 show the received signals after
matched filtering for receiver at a range of 3, 6 and 9 km and depths of 30, 50, 70 and 90m,
respectively. These figures revealed the multipath arrival structures obtained in shallow water.
The multiple reflection effect is obvious. The matched filtered received signals show
variability in the arrival time. The observed time variability in the received signals increase
with increasing range.
Fig.5 The received signals as calculated by BNM (Broadband Normal Modes) model for receiver at a
range of 3 km and depths of 30, 50, 70 and 90m, respectively
Fig.6 The received signals as calculated by BNM model for receiver at a range of 6 km and depths of
30, 50, 70 and 90m, respectively
Fig.7 The received signals as calculated by BNM model for receiver at a range of 9 km and depths of
30, 50, 70 and 90m, respectively
Fig.8 The received signals after matched filtering for receiver at a range of 3 km and depths of 30,
50, 70 and 90m, respectively
4. CONCLUSIONS
The paper has considered the long range shallow water hydroacoustic channel. Chirp
signals have been reviewed in relation to underwater acoustic channel. The multipath effect in
a shallow water channel has been discussed. The received chirp pulses were all matched
filtered showing variability in arrival time and revealing the multipath structure. The results
from numerical experiments clearly show temporal variability in the acoustic signals which is
important pieces of information for the design of shallow water communication systems.
REFERENCES
[1] W.S.Burdic, Underwater acoustic system analysis, Peninsula Publishing, Los Altos
2002.
[2] F.Bini-Verona, P.L.Nielsen, F.B.Jensen, Efficient modeling of broadband propagation
in shallow water, 4th European Conference on Underwater Acoustics, Rome, 21-25
September 1998.
[3] J.P.Hoffman, J.D.Penrose, D.R.McMahon, Acoustic propagation prediction in shallow
water, Australian Acoustical Society Conference, Joondalup, Australia, 15-17
November 2000.
[4] A. Plaisant et.al., Prosim - User Manual of the Prosim Software Package, Annex to the
Final Technical Report, Thomson Marconi Sonar, Nice, 1999.
[5] F.Bini-Verona, P.L.Nielsen, F.B.Jensen, PROSIM broadband normal-mode model. A
user’s guide, SACLANTCEN SM-358, La Spezia, 1999.
[6] M.B.Porter, The KRAKEN Normal Mode Program, SACLANT Undersea Research
Centre, May 17, 2001.
[7] S.Van Dalen, D.G. Simons, R. McHugh, J.F.E. Jackson, Validation of a broadband
normal-mode propagation model using shallow water experimental data, 4th European
Conference on Underwater Acoustics, Rome, 21-25 September 1998.
[8] http://www.saclantc.nato.int/mast/prosim/software/software.html.

Acoustic Communications in Shallow Waters: Sławomir Jastrzębski

Uploaded by

Copyright:

Available Formats

Acoustic Communications in Shallow Waters: Sławomir Jastrzębski

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Acoustic Communications in Shallow Waters: Sławomir Jastrzębski

Uploaded by

Copyright:

Available Formats

ACOUSTIC COMMUNICATIONS IN SHALLOW WATERS

University of Technology and Agriculture in Bydgoszcz

The shallow water channel is an environment which is of particular interest to many

1. BROADBAND NORMAL MODE MODEL

Fig.1 Model of the range-independent environment

2. PULSE COMPRESSION TECHNIQUES

In case of matched filtering:

Fig.2 Transmitted chirp pulse in the time and frequency domain

Fig.3 The model of shallow water hydroacoustic channel

Fig.4 Sound speed profile for tested underwater acoustic channel

Tab.1 The shallow water sound speed profile

Depth [m] Sound speed [m/s]

You might also like