Autonomous Underwater Vehicle Navigation 14.1 Sensors 343
Part B | 14.1
size of proprioceptive navigation sensors have reduced
dramatically, making them available to smaller and
lower-cost AUV systems. The performance claims of
better than 0:1% of distance traveled have been made in
the literature [14.10].
Regardless of sensor cost and quality, the problem
with exclusive reliance on proprioceptive sensing is that
the position error increases without bound as the distance traveled by the vehicle increases. The rate of
increase will be a function of ocean currents, the vehicle speed, and the quality of dead reckoning sensors.
If a vehicle can surface, then GPS can be used for a position fix. Indeed, many AUVs have demonstrated this
capability. However, frequent surfacing is impractical
for deep-water missions and is undesirable for other
types of missions.
In acoustic navigation, transponders serve as beacons to constrain INS/DR error growth without the need
for resurfacing. Two types of systems have been primarily employed [14.11–13]: LBL and ultra-short baseline
(USBL). Both systems employ external transducers or
transducer arrays as aids to navigation. Acoustic navigation is a well-established and widely used technique.
However, in littoral waters, a number of sources of error
occur. These include: multipath, drop-outs, fading, and
reverberation. Further development is required to attain
better outlier rejection and improve robustness for autonomous operations [14.14]. A longer term goal is to
fully integrate navigation, communications, and tomography [14.15].
For many applications, deployment of acoustic beacons is undesirable, and map-based navigation presents
an alternative. If an accurate a priori map of the environment is available, one approach to globally-referenced
position estimation is to use measurements of geophysical parameters, such as bathymetry, magnetic field, or
gravitational anomalies [14.16, 17]. In practice, an upto-date, high-quality map may be unavailable in the
operating area of interest. This motivates research into
the problem of SLAM, in which the goal is for the AUV
to build a map of its environment and to use that map to
navigate in real time.
The structure of this chapter is as follows. Section 14.1 describes available sensors for measuring
depth, heading, velocity, acceleration, acoustic range
and bearing, and GPS. Section 14.2 describes the main
navigation algorithms including DR and inertial navigation, LBL and USBL navigation, geophysical and
map-based navigation, and cooperative localization.
Section 14.3 summarizes the performance of several
typical types of AUV systems in use today, and provides an outlook for future developments in this area.
14.1 Sensors
14.1.1 Depth
All submersible vehicles are outfitted with a pressure sensor [14.18] that allows an accurate determination of absolute depth using the known properties
of sea water [14.19]. As a result, all other underwater
navigation systems are only used to resolve the twodimensional (2-D) position, (i. e., longitude and latitude) and underwater-vehicle-related localization problems are typically stated in 2-D. Quartz crystal pressure
sensors can typically attain accuracies of 0:01% or
better, but achieving full accuracy requires careful calibration and thermal compensation [14.1].
14.1.2 Compass
Like the pressure sensor, a compass is part of the
basic navigation sensor suite of most underwater vehicles, as it is an inexpensive and low-power device.
It provides the three-dimensional (3-D) vector of the
local magnetic field. Before computing the heading
from the magnetic field vector, it is necessary to carefully calibrate the compass each time the vehicle’s
area of operation changes, as the difference between
the orientation of the 3-D magnetic field vector and
the direction of true north (called variation) depends
on the geographic location. In addition to the spatially slow variation, there are highly localized magnetic
anomalies. The compass output is also affected by
its position in the vehicle as electrical currents create magnetic fields which cannot be discerned from
the earth’s magnetic field. On-line compass calibration algorithms [14.3, 20] can substantially improve
performance.
14.1.3 Gyroscopes
Gyroscopic systems measure changes in vehicle orientation by exploiting physical laws that have predictable
effects under rotation [14.21]. Available sensors include mechanical gyrocompasses, ring laser gyroscope
(RLG), fiber optic gyroscope (FOG), dynamically
tuned gyroscope (DTG) and micro-electromechanical
(MEMS) devices [14.22]. Optical gyroscopes have become popular as an accurate angular rate sensor, and
operate on the SAGNAC principle [14.23]. However,
sensible DR navigation requires high gyroscope performance increasing the associated costs significantly.
Part B | 14.1
size of proprioceptive navigation sensors have reduced
dramatically, making them available to smaller and
lower-cost AUV systems. The performance claims of
better than 0:1% of distance traveled have been made in
the literature [14.10].
Regardless of sensor cost and quality, the problem
with exclusive reliance on proprioceptive sensing is that
the position error increases without bound as the distance traveled by the vehicle increases. The rate of
increase will be a function of ocean currents, the vehicle speed, and the quality of dead reckoning sensors.
If a vehicle can surface, then GPS can be used for a position fix. Indeed, many AUVs have demonstrated this
capability. However, frequent surfacing is impractical
for deep-water missions and is undesirable for other
types of missions.
In acoustic navigation, transponders serve as beacons to constrain INS/DR error growth without the need
for resurfacing. Two types of systems have been primarily employed [14.11–13]: LBL and ultra-short baseline
(USBL). Both systems employ external transducers or
transducer arrays as aids to navigation. Acoustic navigation is a well-established and widely used technique.
However, in littoral waters, a number of sources of error
occur. These include: multipath, drop-outs, fading, and
reverberation. Further development is required to attain
better outlier rejection and improve robustness for autonomous operations [14.14]. A longer term goal is to
fully integrate navigation, communications, and tomography [14.15].
For many applications, deployment of acoustic beacons is undesirable, and map-based navigation presents
an alternative. If an accurate a priori map of the environment is available, one approach to globally-referenced
position estimation is to use measurements of geophysical parameters, such as bathymetry, magnetic field, or
gravitational anomalies [14.16, 17]. In practice, an upto-date, high-quality map may be unavailable in the
operating area of interest. This motivates research into
the problem of SLAM, in which the goal is for the AUV
to build a map of its environment and to use that map to
navigate in real time.
The structure of this chapter is as follows. Section 14.1 describes available sensors for measuring
depth, heading, velocity, acceleration, acoustic range
and bearing, and GPS. Section 14.2 describes the main
navigation algorithms including DR and inertial navigation, LBL and USBL navigation, geophysical and
map-based navigation, and cooperative localization.
Section 14.3 summarizes the performance of several
typical types of AUV systems in use today, and provides an outlook for future developments in this area.
14.1 Sensors
14.1.1 Depth
All submersible vehicles are outfitted with a pressure sensor [14.18] that allows an accurate determination of absolute depth using the known properties
of sea water [14.19]. As a result, all other underwater
navigation systems are only used to resolve the twodimensional (2-D) position, (i. e., longitude and latitude) and underwater-vehicle-related localization problems are typically stated in 2-D. Quartz crystal pressure
sensors can typically attain accuracies of 0:01% or
better, but achieving full accuracy requires careful calibration and thermal compensation [14.1].
14.1.2 Compass
Like the pressure sensor, a compass is part of the
basic navigation sensor suite of most underwater vehicles, as it is an inexpensive and low-power device.
It provides the three-dimensional (3-D) vector of the
local magnetic field. Before computing the heading
from the magnetic field vector, it is necessary to carefully calibrate the compass each time the vehicle’s
area of operation changes, as the difference between
the orientation of the 3-D magnetic field vector and
the direction of true north (called variation) depends
on the geographic location. In addition to the spatially slow variation, there are highly localized magnetic
anomalies. The compass output is also affected by
its position in the vehicle as electrical currents create magnetic fields which cannot be discerned from
the earth’s magnetic field. On-line compass calibration algorithms [14.3, 20] can substantially improve
performance.
14.1.3 Gyroscopes
Gyroscopic systems measure changes in vehicle orientation by exploiting physical laws that have predictable
effects under rotation [14.21]. Available sensors include mechanical gyrocompasses, ring laser gyroscope
(RLG), fiber optic gyroscope (FOG), dynamically
tuned gyroscope (DTG) and micro-electromechanical
(MEMS) devices [14.22]. Optical gyroscopes have become popular as an accurate angular rate sensor, and
operate on the SAGNAC principle [14.23]. However,
sensible DR navigation requires high gyroscope performance increasing the associated costs significantly.