IEEE JOURNAL OF OCEANIC ENGINEERING, VOL. 35, NO. 4, OCTOBER 2010 877
Peer-Reviewed Technical Communication
Modeling Bottom Reverberation for Sonar Sensor Motion
Jinyun Ren, Student Member, IEEE, and Rodney G. Vaughan, Fellow, IEEE
Abstract—Simplified signal models of the bottom reverberation
are widely used in the sonar literature, but the simplifications
themselves are seldom discussed. In this paper, the formulation of
the joint modeling of signal propagation, bottom scattering, and
reception is reviewed, and an emphasis is placed on the assump-
tions which lead to model simplification. The interest extends to
modeling for sensor motion compensation algorithm development,
and it is in this context that an in-depth understanding of the
assumptions and simplifications becomes essential. The formula-
tions help pave the way for the derivation of motion compensation
algorithms. One such algorithm is presented in the companion
paper.
Index Terms—Bottom reverberation, motion compensation,
scattering and propagation modeling, sonar sensor motion.
I. INTRODUCTION
A
major problem in active sonar systems is reverberation,
which is volume scattering from the water body and sur-
face scattering from the water surface or the water bottom (al-
ternatively referred to as the seafloor) [1], [2]. In this paper, we
address the scattering contributions from the water bottom. The
bottom reverberation is an undesired signal in some applications
such as target detection in shallow water, while it is the wanted
signal in seafloor mapping.
In conventional applications such as sonar depth sounders,
the amplitude or power of bottom reverberation is of interest.
However, with the introduction of sonar arrays, both the phase
and amplitude of the received bottom reverberation become im-
portant to the corresponding algorithms. For example, swath
bathymetric sidescan sonars calculate depth by using the phase
difference between signals received from different elements of
a sonar array [3]. Here, the time of flight gives range, and the
receiving array gives an angle of arrival, and these parameters
allow the depth to be calculated. In these applications, there is
seldom ground truth available, so the raw and processed data
must be interpreted carefully. Consequently, recent theoretical
performance analysis of this kind of sonar relates to character-
Manuscript received April 04, 2009; revised July 02, 2010; accepted
September 15, 2010. Date of publication November 11, 2010; date of current
version November 30, 2010.
Associate Editor: D. Knobles.
The authors are with the School of Engineering Science, Simon Fraser
University, Burnaby, BC V5A 1S6 Canada (e-mail: jren@alumni.sfu.ca;
rodney_vaughan@sfu.ca).
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/JOE.2010.2079591
izing estimation accuracy, e.g., [4] and [5]. Similarly, for syn-
thetic aperture sonar (SAS), the signal processing algorithms
(including vehicle motion compensation and bottom imaging)
and corresponding performance analysis rely on the phase as
well as the amplitude of the received signal [6]–[9].
To develop sonar signal processing algorithms and conduct
theoretical performance analysis, the modeling of physical
mechanisms of bottom reverberation should be thoroughly
examined. A simplified signal model of bottom reverberation,
defined from a plane (i.e., 2-D only) and assuming the bottom
to be like a line (i.e., 1-D), is adopted by many researchers [4],
[5], [10]. For example, in [4] and [5], this model acts as a start
point for deriving the correlation function of reverberation sig-
nals received by different elements of a sonar sensor array, and
this correlation function is further used to analyze the accuracy
of swath bathymetric sonars. A more general bottom reverber-
ation signal model (considering the bottom as a 2-D surface)
is adopted in deriving the space-time correlation function of
signals scattered from the bottom for correlation sonars in [11].
This resulting correlation is also the key for estimating the
displacement of a 3-D sonar array. However, these references
do not elaborate the simplifications or their justification; or
the interplay between the sonar mechanisms, their modeling,
their assumptions, and the algorithms used. The modeling of
bottom reverberation becomes important when discussing the
sonar signal itself, in particular its phase, because the algorithm
performance is typically derived from a model of the complex
transmission characteristics.
The bottom reverberation physics is modeled using rough sur-
face scattering theory (e.g., [12]–[15]), pioneered in electro-
magnetic wave research (e.g., [16]–[21]). These texts mainly
discuss scattered power, rather than complex amplitudes, in-
cluding the phase features of scattered signals. To the authors’
knowledge, there is no concise discussion of the joint modeling
of signal propagation, bottom scattering, and reception. In this
paper, the bottom reverberation modeling for a stationary sonar
is discussed in detail. The model derives from the principles of
propagation, and includes all the steps to the known formula
in the sonar literature. The important assumptions and simpli-
fications, which are not previously discussed in the literature,
are significant for algorithms and performance analysis. The
assumptions are particularly important for developing motion
compensation algorithms. It is shown how it is possible to esti-
mate the sensor motion by comparing two signals received from
the same sensor element but at different times. Using the known,
simplified model for monostatic sonar, an analysis of the effects
of small sensor motion is presented. The assumptions which
0364-9059/$26.00 © 2010 IEEE