Contents lists available at ScienceDirect
Ocean Engineering
journal homepage: www.elsevier.com/locate/oceaneng
A customized H-infinity algorithm for underwater navigation system: With
experimental evaluation
Mehdi Emami
a
, Mohammad Reza Taban
b,
⁎
a
Department of Electrical and Computer Engineering, Yazd University, Iran
b
Department of Electrical and Computer Engineering, Isfahan University of Technology, Isfahan 84156-83111, Iran
ARTICLE INFO
Keywords:
DVL
INS
Underwater Navigation System
H-infinity filter
ABSTRACT
In this paper, an integrated navigation system based on H-infinity filter is proposed for improving the
performance of underwater navigation system. The proposed navigation system consists of a package with a
triad of accelerometers and an Attitude and Heading Reference System (AHRS) as the main system and a
Doppler Velocity Log (DVL) as an auxiliary system. In this system, the dynamic and measurement models are
defined in such a way by which a H-infinity based integrated navigation system is derived. In addition, according
to the specific dynamic of discussed vehicle, a scheme is proposed for estimating the performance bound
parameter of H-infinity filter. The proposed system becomes robust versus the system modeling errors and noise
uncertainty in comparison with the conventional Kalman Filter (KF). Furthermore, it can be effectively decrease
the effect of the DVL's outlier data. The performance of the proposed system is evaluated through several sea
tests using an Autonomous Underwater Vehicle (AUV). Furthermore, the conventional Dead Reckoning (DR)
and KF algorithms are utilized for comparison. The experimental results demonstrate the superior performance
of the proposed system compared with the conventional algorithms.
1. Introduction
The Dead Reckoning (DR) algorithm is the most common method
for estimating the position of an underwater vehicle (Brokloff, 1997).
In this algorithm, the current position is calculated by knowing the
previous position, and the measured velocity and heading (Groves,
2008). In underwater navigation, the velocity and orientation of the
underwater vehicle are usually measured by Doppler Velocity Log
(DVL) and Attitude and Heading Reference System (AHRS), respec-
tively (Kinsey et al., 2006). DVL is an acoustic sensor which can
measure the vehicle's velocity relative to the water (water tracking) or
sea bottom (bottom tracking) through emitting signals to the bottom of
the sea and using Doppler shift of the returned signals (Grenon et al.,
2001; Farrell, 2008). In order to have an accurate positioning, the DVL
must measure the vehicle's velocity relative to the sea bottom.
However, DVL bottom tracking does not always occur. Some factors
such as acoustic noises, the type of the seabed and exceeding of the
operational range may lead to outlier data or drop out in the DVL's
output. The main shortcoming of the DR algorithm is the bias of the
velocity and heading measurement which leads to an accumulative
error proportional to time in position estimate (Jalving et al., 2003).
Outlier data and drop out in DVL's measurements reduce the DR
accuracy.
Strap-Down Inertial Navigation System (SDINS) is another algo-
rithm for underwater navigation in which an IMU (Inertial
Measurement Unit) is used for calculating the vehicle's position,
velocity and orientation according to and the DR principle (Titterton
and Weston, 2004). The IMU consists of three orthogonal acceler-
ometers and three orthogonal gyroscopes which measures the accel-
eration and angular rate signals in x, y and z directions of the body
frame. The INS has also accumulative error in position due to bias
errors in the accelerometers and gyroscopes and sequential integra-
tions of these signals (Wang et al., 2015; Shabani et al., 2013).
In order to reduce the error growth of the SDINS, the auxiliary
sensors must be utilized Ali et al. (2012), Gao et al. (2014). Global
Positioning System (GPS), DVL, depth meter, AHRS and Acoustic
Positioning Systems (APS) can aid the SDINS (Paull et al., 2014). Due
to unavailability of GPS signals under the water, it is not possible to use
it in deep waters applications (Grenon et al., 2001; Allotta et al., 2016).
The APS utilization is limited due to high cost, time-consuming and
difficult position calibration of transponders, and limitation of its
operational range (Stutters et al., 2008; Lee and Jun, 2007; Lee
et al., 2007; Chen, 2008; Caiti et al., 2014). Nevertheless, DVL, depth
meter and AHRS can be suitable sensors for long-term underwater
http://dx.doi.org/10.1016/j.oceaneng.2016.12.011
Received 9 March 2016; Received in revised form 4 September 2016; Accepted 18 December 2016
⁎
Corresponding author.
E-mail addresses: m.emami@stu.yazd.ac.ir (M. Emami), mrtaban@cc.iut.ac.ir (M.R. Taban).
Ocean Engineering 130 (2017) 611–619
Available online 28 December 2016
0029-8018/ © 2016 Elsevier Ltd. All rights reserved.
MARK