Ocean Engineering 110 (2015) 303–313

Contents lists available at ScienceDirect

Ocean Engineering
journal homepage: www.elsevier.com/locate/oceaneng

A survey on path planning for persistent autonomy of autonomous

underwater vehicles
Zheng Zeng a,b,n, Lian Lian a, Karl Sammut b, Fangpo He b, Youhong Tang b,
Andrew Lammas b
a
Institute of Oceanology, Shanghai Jiao Tong University, Shanghai 200240, China
b
Centre for Maritime Engineering, Control and Imaging, School of Computer Science, Engineering and Mathematics, Flinders University, SA, Australia

art ic l e i nf o a b s t r a c t

Article history: Recently, there has been growing interest in developing Autonomous Underwater Vehicle (AUV) to
Received 16 March 2015 operate for longer mission durations as well as with higher levels of autonomy. This paper describes the
Received in revised form current state of the art in methodologies to enable long range AUVs and provides a detailed literature
1 October 2015
review of some existing AUVs characterizing their operation endurance. Path planning is identified as the
Accepted 5 October 2015
Available online 8 November 2015
core and crucial components to improve AUV persistence. The aims of path optimization, path re-
planning adapting to the dynamic environments, and cooperative path planning of multiple AUVs have
Keywords: received much attention from the research community. This paper presents a review of the main
Autonomous underwater vehicle research works focusing on these three technology areas. The main objective of this paper is to present a
Path planning
comprehensive survey of shape and properties of the path and optimization techniques for path plan-
Persistent autonomy
ning. These techniques and algorithms have been classified into different classes and their assumptions
Path optimization
and drawbacks have been discussed. Finally, the paper discusses the AUV literature in general and
highlights challenges that need to be addressed in developing AUVs with advanced autonomy and
capable of operating for longer mission durations.
& 2015 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
1.1. Motivations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
2. Operation endurance levels for AUVs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
2.1. AUVs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
3. Path planning problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
3.1. Path optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
3.1.1. Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
3.1.2. Energy consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
3.1.3. Travel time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
3.2. Path re-planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
3.3. Cooperative path planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
4. Path shape and properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308
5. Optimization techniques for path planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
5.1. Graph search shemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
5.2. Fast Marching and Level Set Methods (FM & LSM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
5.3. Artificial Potential Field (APF) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
5.4. Rapidly-exploring Random Trees (RRT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
5.5. Evolutionary algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
6. Open problems and research challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
6.1. Path planning in complex ocean environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311

n
Corresponding author at: Institute of Oceanology, Shanghai Jiao Tong University, 800 Dong Chuan Road, Minhang, Shanghai, China. Tel.: þ86 15221545687.
E-mail address: zheng.zeng@flinders.edu.au (Z. Zeng).

http://dx.doi.org/10.1016/j.oceaneng.2015.10.007
0029-8018/& 2015 Elsevier Ltd. All rights reserved.
304
6.2. On-line path replanning in spatiotemporal environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.3. Path planning for multi-AUVs rendezvous . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1. Introduction
Autonomous Underwater Vehicles (AUVs) are a class of submerged
marine robots using various enabling technologies to navigate and
perform various tasks. AUVs have a variety of military, scientific
research and commercial applications (Westwood, 2010). Military
applications include surveillance and reconnaissance, anti-submarine
warfare, payload delivery, time-critical strike, harbor protection etc.,
AUVs are useful for oceanography, littoral ocean floor mapping, water
profile sampling and other types of scientific research, AUVs are also
increasingly used in the oil and gas market, due to their capability to
operate very close to the seabed or in close proximity to industrial
structures like those used by offshore industries.
However, the actual autonomy of contemporary AUVs is limited
in many ways restricting their potential uses. Further advances in
AUV autonomy will enable new operations, such as executing very
long endurance missions with a minimum of supervision in
unknown, dynamic and hostile environments.
The level of autonomy achieved by AUVs is chiefly determined
by their performance in path planning and re-planning. A path
planner should be capable of finding a trajectory that safely leads
the AUV from its initial or current position to its destination and
optimize a certain objective function, such as time or energy
consumption. The environment usually contain time-varying cur-
rents and obstacles which might not be fully characterized at the
start of a mission. Obstacles may be detected as the AUV moves
through the environment, rendezvous location may change over
time. A special case of non-stationary rendezvous target, such as a
surface vehicle, may have moved during the operation. The path
planner should thus have the capability to adapt to changing
ocean environment, mission goals and system status.
1.1. Motivations
Many of current AUV applications consume much time lasting from
days to weeks, and cover large areas of hundreds to thousands of square
kilometers. Oceanographic processes, such as currents show great
variability over such large expanses and will evolve over such durations.
In most existing applications, AUVs are typically deployed from surface
vessels with support personnel for deployment, piloting and recovery.
During the course of the mission, the support vessel will shadow the
AUV and provide much of the higher level decision making processes
needed to deal with changes of oceanographic processes. The cost of
keeping the support vessel on standby is generally by far the most
significant component of the mission cost (Furlong et al., 2012).
Recently, there has been growing interest in developing long range
AUVs with increased autonomy to conduct science missions over longer
periods without supervision (Hobson et al., 2012), thereby reducing
mission costs and extending their applicability.
Two developments have been explored to increase the range
and endurance of the long-range AUV. The first development is to
increase the vehicle storage density-which indicates the capability
to store more energy and sensors per unit volume. The second
development is to increase the efficiency of the vehicle. These
developments are illustrated in Fig. 1.
Compact battery technology capable of storing more electricity
is one way to increase the vehicle storage density and capability.
Most AUVs in use today are powered by rechargeable batteries, such
Z. Zeng et al. / Ocean Engineering 110 (2015) 303–313
311
311
311
311
as lithium ion, lithium polymer, nickel metal hydride etc. Some
vehicles use primary batteries which provide perhaps twice the
endurance, but at a substantial extra cost per mission. A few of the
larger vehicles are powered by aluminum based semi-fuel cells,
however, these require substantial maintenance, require expensive
refills and produce waste product that must be handled safely. An
emerging trend is to combine different battery and power systems
with super-capacitors (Abu Sharkh and Griffiths, 2002; Hasvold et
al., 2006). Another way to increase the vehicle's storage density
capacity is to decrease the payload size (Taylor and Wilby, 2011),
application of nanotechnology for developing sensor equipment
holds potential to develop highly sophisticated underwater vehicles
with reduced payload sizes and power consumption.
Research to increase the efficiency of the AUVs can be cate-
gorised along three directions. One direction is looking at reducing
the AUV power consumption, including both propulsion power and
hotel load (Hobson et al., 2012). Power for most AUVs is generated
from power stored in on-board batteries. Propulsion power can be
reduced through more efficient motion. Hotel load is defined as the
power for on-board instrumentation, guidance, computers and
communication devices, and has steadily reduced through
advancements in electronic systems. A second direction is looking
at acheving advances in mechanical design, particularly the vehicle
profile and surface design to reduce vehicle drag (Joung et al., 2009;
2012). A good example of this category is Tethys, its hull, motor and
propeller were computer designed and tested to minimize drag and
maximize efficiency of propulsion. The third major direction is
increase the autonomy of vehicle (Hobson et al., 2012). Mission
scheduling and path planning hold tremendous potential to enable
long range operations. A vehicle could be launched from shore
where upon a path planning system could be used to generate a
trajectory that exploits the ocean energy taking use of the favorable
currents to propel the vehicle, and lead the vehicle to a remote
work site, perform a survey, and then return to shore completely on
its own. This will greatly diminish the costs for AUV operations
since there is no need for a support vessel and vehicle operations
can be monitored on-shore, thereby enhancing the affordability of
AUVs to science and industry.
The rest of the paper is organized as follows: Section 2 lists
some existing AUVs categorize them into three classes based on
their operation endurance. Section 3 introduces the AUV path
planning problem, including the aims of path optimization, online
path re-planning, and cooperative path planning of multiple AUVs.
Sections 4 and 5 provide a comprehensive survey of major works
focusing on shape and properties of the path and optimization
Long range AUV enabling technologies
Improved Efficiency Improved Storage Density
Mechanical Autonomy Power consumption Battery Payload
Drag reduction Mission schedule Propulsion power
Skin friction Path Planning Hotel load
Fig. 1. Methodologies to enable long range AUVs.
 Z. Zeng et al. / Ocean Engineering 110 (2015) 303–313 305

techniques for path planning with discussion of their assumptions
and drawbacks, respectively. Outline some open problems and
research challenges to be addressed in Section 6.
2. Operation endurance levels for AUVs
Improving persistent presence is at the forefront in AUV
research. AUVs are increasingly expected to operate for longer
mission durations as well as with higher levels of autonomy and
sensor capability to make efficient use of the vehicle and to reduce
support vessel costs.
2.1. AUVs
A review of the existing AUVs suggests that they can be cate-
gorized into three broad classes based on their operation endurance.
 Category I: Long-range AUV. Long-range can travel for several
thousand of kilometers to conduct science missions over peri-
ods of weeks or even month. Autosub Long Range is designed by
the National Oceanography Center (NOC) with a target range of
6000 km with over six months endurance. This range is
achieved with slow operation velocity and efficient use of the
onboard energy (Furlong et al., 2012). Another good example of
this category is the Reliant AUV developed by the US Naval
Research Laboratory (NRL), which executed a record setting
endurance of 109 h traveling nearly 315 miles, at a depth of
10 m and an average speed of 2.5 knots through busy water-
ways and strong currents (2013).
 Category II: Medium-range AUV are autonomous submersible
platforms which have an endurance of more than 72 h with
nominal load. Some examples include the Bluefin-12 (US),
Autosub6000 (UK), Hugin 3000 (Norwegian) and the Remus
600 (US). The main advantage of these platforms is their
mission endurance which allows them to carry various heavy
and high-quality navigation and mission sensors. These plat-
forms are generally supported with a research vessel, so that
their batteries can be regularly replaced and the vehicle can be
maintained.
 Category III: Small-range AUV are autonomous submersible
platforms which have an endurance of less than 24 h. AUVs
such as the Bluefin-9, Remus 6000, Teledyne Gavia, and MUNIN
AUV and others, fall into this category. Platforms of this class
have less endurance than those of Category II, but they feature
low cost, greater mobility and more commercial availability
than other catergories.
The main feature of Categories I – long range AUVs is their
significant endurance and range. It is a sign of the increasing
maturity of autonomous vehicles that users are now demanding
longer mission durations as well as higher levels of autonomy and
sensor capability to make most effective use of the technology.
AUVs of Categories I and II are generally suitable for open sea
applications and require a suitable facility for launch and recovery.
Most short-range AUVs of Categories III can be applied open sea as
well as constrained environment like coastal, river and harbor, and
can be launched by hand.(Tables 1–3)

Table 1
Category I-Long Range AUV (Endurance Z72 h).

Name Manufacture Size(LxWxH) Weight Depth Speed Energy Endurance Figure

Autosub long National oceano- 3.60 m  0.90 m  0.90 m 660 kg 6000 m 0.40 m/s 18.9 KW h 4400 h
range (Fur- graphy centre
long et al.,
2012)

Tethys (Hob- MBARI 2.30 m  0.31 m  0.31 m 110 kg 200 m 0.50 m/s Information 740 h
son et al., not available
2012)

Reliant U.S. Naval 6.1 m  0.53 m  0.53 m 612 kg 4500 m 1.28 m/s 40 KWh 109 h
(Navy'- research labora-
Mine-Hun- tory's (NRL) þ
ter' AUV bluefin robotics
Sets Mis-
sion
Endurance
Record)
306 Z. Zeng et al. / Ocean Engineering 110 (2015) 303–313

Table 2
Category II-Medium Range AUV (24 h o Enduranceo72 h).

Name Manufacture Size(LxWxH) Weight Depth Speed Energy Endurance Figure

Autosub6000 National ocea- 5.50 m  0.90 m  0.90 m 2000 kg 6000 m 1.00 m/s 4.5 KW h 70 h
(Griffiths and nography centre
McPhail, 2011)

Remus 600 (Sto- Hydroid 4.27 m  0.32 m  0.32 m 326 kg 600 m 1.50 m/s 5.2 KW h 50 h
key et al.,
2005)
Bluefin-12 (Tay- Bluefin robotics 3.77 m  0.32 m  0.32 m 204 kg 200 m 1.50 m/s 4.5 KW h 26 h
lor and Wilby,
2011)

Bluefin-21 (Blue- bluefin robotics 4.10 m  0.53 m  0.53 m 525 kg 4500 m 1.54 m/s 13.5 KW h 25 h
fin-21 Product
Sheet - Bluefin
Robotics)
Hugin 1000 Kongsberg 4.70 m  0.75 m  0.75 m 850 kg 3000 m 2.05 m/s 15 KW h 24 h
(Hagen et al., maritime
2003)

Table 3
Category III. Short Range AUV (Endurance r 24 h).

Name Manufacture Size (LxWxH) Weight Depth Speed Energy Endurance Figure

MUNIN (Autono- Kongsberg 0.30 m  0.03 m  0.03 m 300 kg 1500 m 2.05 m/s 5 KW h 22 h
mous under- maritime
water vehicle -
MUNIN AUV)
Bluefin-9 (Bluefin- Bluefin robotics 1.65 m  0.24 m  0.24 m 50 kg 200 m 1.52 m/s 1.5 KW h 12 h
9 Product Sheet
- Bluefin
Robotics)

Iver2 (Incze, 2011) Ocean server 0.13 m  0.01 m  0.01 m 19 kg 100 m 1.29 m/s 0.6 KW h 12 h
technology

Remus 100 Hydroid 1.84 m  0.19 m  0.19 m 45 kg 100 m 2.3 m/s 1 KW h 10 h

(Nygaard, 2014)

Teledyne Gavia Gavia defence 1.80 m  0.20 m  0.30 m 49 kg 1000 m 1.00 m/s 1.2 KW h 7h

3. Path planning problem
Path planning for AUVs in the ocean has become crucial for
many applications, ranging from security and acoustic surveil-
lance, to collection of ocean data at specific locations, for ocean
prediction and monitoring. This paper reviews several problems
associated with path planning and re-planning of AUV missions
through cluttered ocean environments and contributes to the
desired increase in AUV autonomy.
The following is a general description of the AUV path planning
problem:
Given the present state of a single AUV or team of AUVs and
information of the ocean environment, compute a trajectory
towards a desired goal state that optimizes a certain objective
function while respecting the properties of the vehicle(s) and
avoiding collisions with obstacles (Kuwata et al., 2005).
3.1. Path optimization
In general, the path planning problem aims to optimize at least
one of the following aspects of performance:
 Z. Zeng et al. / Ocean Engineering 110 (2015) 303–313
3.1.1. Safety
Traditionally, path planning for AUVs has been related to safety
conditions. The path should be devoid of known obstacles or
hazardous areas. Knowledge of an obstacle's position is usually not
perfect. Obstacles may be detected as the vehicle moves through
the environment or their location may change over time. A special
case of moving obstacles are other intelligent agents with which
collisions should be avoided. The vehicle should thus have the
capability to compute or update its path in real-time. Different
computational methods were employed to plan safe paths for AUVs.
Warren (1990) used artificial potential field (APF) algorithms to
solve the path planning problem. The algorithm generates safe
paths by introducing an artificial potential field around the obsta-
cles preventing vehicles from collision. While the use of the APF
cost-terms allows fast convergence and is easy to apply to irregu-
larly shaped obstacles, it is susceptible to local minima. Graph
Search Methods include Dijkstra's algorithm (Dijkstra, 1959), A*
algorithm (Carroll et al., 1992) and the D* algorithm (Ferguson and
Stentz, 2006) have been employed for path planning. In this
methods, a chart or graph is produced showing free space where no
collision will occur and forbidden spaces where a collision will
occur. Based on this graph, a path is selected by piecing together the
free spaces or by tracing around the forbidden spaces. Graph Search
methods are fast to solve low-dimensional problems, however, they
are commonly criticized for their discrete state transitions which
unnaturally constrain the motion of a vehicle to limited directions.
Risk-aware path planners are recently reported in Pereira et al.
(2011) described path-planning methods that make safe operating
of AUVs in high ship and boat traffic areas. This work is extended in
Pereira et al. (2013) by looking at minimizing risk through search in
a time expanded planning environment while taking advantage of
ocean current predictions. While it is good to see that this work use
ocean current predictions instead of quasi-static currents data for
planning AUV paths, however, it still applies A* as the optimization
algorithm which is not suitable for high-demensional problems.
3.1.2. Energy consumption
The problem of AUV mission planning to optimize the energy
cost of traveling in ocean environments with quasi-static current
fields was considered by Alvarez et al. (2004) and Garau et al.
(2009). The developed planning algorithm integrated the currents
map with an evolutionary path planner, providing the path with
minimum energy requirements. In these studies the AUV speed
with respect to the seabed was assumed to be constant. Thus, the
AUV should be able to adapt its speed depending on the current
field, keeping the total speed constant through the planned tra-
jectory. One of main drawbacks of this work is that it makes the
assumption that all the paths are strictly monotonic with respect to
the x-coordinate, while the y-coordinates are set up as the chro-
mosome genes for the evolutionary algorithm. Koay and Chitre
(2013) used A* to find a path for an AUV to navigate in the presence
of currents and obstacles while consuming minimum amount of
propulsion energy. The authors concluded that the expected energy
saving is the largest when the vehicle is allowed to operate at
speeds comparable to the water current while the savings diminish
when the vehicle is required to operate at higher speeds. An
alternate approach to minimum energy continuous path planning
was proposed by Kruger et al. (2007). This work considered vehicle
actuation limits and use a multi-dimensional cost function for
generating energy optimum paths in estuarine environments.
Regarding the optimization algorithm, however, this work only
applied a simple, gradient based approach for proving the concept.
Thus, the successful optimization relies heavily on a careful choice
of path parameterization. The Particle swarm optimization (PSO)
algorithm has also been studied (Witt and Dunbabin, 2008) for
energy conservation by taking advantage of the time-varying ocean
307
currents. Like other evolutionary algorithms, its drawback is that
they may converge to a suboptimal solution within a finite time.
3.1.3. Travel time
Garau et al. (2009) considered a more common situation on
actual AUVs, where the thrust power is usually kept constant
during the mission. Thus, in the usual situation the optimization of
energy consumption agrees with finding the minimum-time path.
This work also used A* search to find the optimal paths, so it will
face the same problem of being computationally expensive to
employ in high-dimensional search spaces. Fast marching method
was applied by Pêtrès et al., (2005), (2007) to obtain the minimum
time path in a given current field. Similar research that applies
wavefront methods are the spatiotemporal wavefront algorithm
described in Thompson et al. (2010) and the sliding wavefront
expansion algorithm described in Soulignac et al. (2008), Sou-
lignac (2011). Level set methods (Lolla et al., 2012) are more
recently introduced for path planning in a currents fields whereby
the time-optimal path of the AUV is generated by solving a particle
tracking equation backward in time. These deterministic methods
have proven to be quite effective in low-dimensional spaces, but
significantly less effective in high-dimensional settings.
3.2. Path re-planning
Despite the advances made through current AUV research
endeavors, the existing or planned AUV systems are far from cap-
able of fully autonomously taking over on-line dynamic and com-
plex operations. An effective fully-autonomous AUV system, should
be capable of reconfiguring itself in response to environmental
changes in order to improve mission performance and optimize the
strategy for detection and measurement of vector/scalar fields and
features of particular interest (adaptive ocean sampling).
Dynamic path planning can be solved by correcting the previous
path solution based solely on new information (such as newly
detected obstacles). One approach by Wzorek and Doherty (2006),
Wzorek et al. (2010) suggest that for those situations that need to be
handled with urgency, only the colliding segments need be cor-
rected. Given the range of a typical AUV sonar sensor and the max-
imum AUV travel speed, there may be sufficient time, however, to
generate a better solution that can take into consideration both new
and previous information. Previous research, such as Jung and Tsio-
tras (2008) and Cowlagi and Tsiotras (2008), also looked at solving
the dynamic path planning problem by using a hierarchical
approach. This approach generates a solution with greater accuracy
locally where the vehicle's immediate reaction to an obstacle or a
threat is needed. However, it may not be suitable in dynamic
environments where the current field is continuously changing
during the mission. Another approach based on Rapidly-exploring
Random Trees (RRT) (Kuffner and La Valle, 2000) pro-poses a re-
planning strategy called Dynamic RRT (Ferguson et al., 2006; Zucker
et al., 2007), where the RRT is re-paired by pruning newly-invalid
branches and then growing the remaining tree until a new solution is
achieved. RRT strategy is suitable for handling dynamic environ-
ments with moving targets and obstacles, however, it has not been
designed to handle optimal path planning with dynamic currents.
3.3. Cooperative path planning
Recently, there also has been widespread interest in the use of
fleets of autonomous robots cooperating to achieve a desired goal.
Multiple cooperative AUV systems hold great promise for use in large
scale oceanographic surveys, mine counter-measurement missions
and other similar underwater tasks due to better resource and task
allocation. Simultaneous use of multiple vehicles can improve per-
formance, reduce mission time, provide broader or more robust data,
308 Z. Zeng et al. / Ocean Engineering 110 (2015) 303–313
and increase the likelihood of mission success. The ability to ren-
dezvous at a specified location is one of the mission requirments for
multiple cooperative vehicle systems (Tsourdos et al., 2010). The
purpose of rendezvous could be an initial step in formation creation,
data exchange, vehicle recharging, maintenance, or collection. The
literature on cooperative use of multiple vehicles is vast. Specific
attention is given to a number of problems including cooperative
motion control of fleets of autonomous vehicles (Fiorelli et al., 2006),
and cooperative vehicle localization and navigation (Fallon
et al., 2010; Knudson and Tumer, 2011).
Versatile path planning system is of crucial importance to
success and efficient rendezvous. However, previous work is
dominated by computing optimal trajectories for single-vehicle
(Alvarez et al., 2004; Rao and Williams, 2009). There is still con-
siderable work to be done to develop advanced methods for
cooperative multiple marine vehicle path planning that explicitly
address simultaneous rendezvous. One challenge that arises is that
it may be very difficult and energy consuming for the early arrived
vehicle to wait for the other vehicles to arrive, especially if there is
a bad weather situation or a temporary adverse ocean current
condition in the region of rendezvous.
4. Path shape and properties
The shape and properties of the path have a direct influence on
the path planning system. Generally, the path-planning algorithm
will first define a number of ordered waypoints on the operation
space which will have to be connected sequentially so as to form the
path. Connecting the waypoints can be achieved in many different
ways, with each one having its own advantages and drawbacks (see
Table 4). However three main categories are distinguished here:
 Category 1: Straight line paths
In many early applications (Kanayama and Yuta, 1988) the path
is defined as the sum of the successive straight lines that connect
these waypoints. Due to physical constraints though, it is not
possible for a vehicle to achieve a smooth transition between two
straight lines because such a path generally has a discontinuous
first derivative at the locations of the waypoints.
 Category 2: Dubins and similar paths.
Table 4
Literature review of path shape and properties.
Path shapes Reference
Straight lines Kwak et al., 1990
Dubins Straight lines þcircular arcs Dubins, 1957
Straight lines þClothoid arcs Fraichard and Scheuer,
2004
Straight lines þPythagorean Bruyninckx and Rey-
hodographs naerts, 1997
Straight lines þFermat’s spiral arc Dahl, 2013
Piecewise polynomial and Piecewise-constant polynomials Qu et al., 2004
splines
Natural splines Fossen, 2011
Cubic hermite spline Sprunk, 2008
B-spline Jolly et al., 2009
Continuing the first category, a simple and intuitive way of pro-
ducing continuous paths without sharp corners can be achieved by
inscribing a circle between two lines to form a curved path. Dubins
(1957), assumed that for a vehicle moving at a constant speed, the
shortest possible path that meets a maximum curvature bound
between a start and end position with their predefined orientation,
respectively will consists of at most three pieces, each of which is
either a straight line or an arc of a circle of appropriate radius. Dubins
Paths consisting of straight lines and circular arcs have been studied
extensively in (Reeds and Shepp, 1990; Chitsaz and LaValle, 2007;
Hota and Ghose, 2013), Moreover, Techy and Woolsey (2009),
Bakolas and Tsiotras (2010), (2013), deal with the problem of finding
a Dubins path for a vehicle that moves in a constant drifted (can be
wind or current for aerial or marine vehicles, respectively). In this
case, it is not always possible to find a Dubins path. The main dis-
advantage of the Dubins path is the curvature discontinuity which
occurs at the meeting points of two consecutive path segments, for
more details see Tsourdos et al. (2010).
The continuous circular arcs can also be produced by using other
methods. Three of the most popular approaches are the clothoid
arcs (Fraichard and Scheuer, 2004), the Pythagorean Hodographs
(Bruyninckx and Reynaerts, 1997) and the Fermat's spiral arc (Lek-
kas and Fossen, 2014). The clothoid is useful in path-planning
applications due to its property of having its curvature change lin-
early with arc-length. This notion can be extended to three
dimensions, and is consequently also true for the torsion (Harary
and Tal, 2012). Pythagorean hodographs employ a polynomial of the
fifth degree in order to produce a closed-form solution that gives a
flexible path with velocity continuity. Further work has been done
in (Bruyninckx and Reynaerts, 1997) so as to guarantee that the
curvature constraints are satisfied as well. The main drawback of
these approaches is the increased computational cost which can be
a heavy burden when implementing them in, for instance, on-line
AUV path planning. A more detailed comparison of Dubins paths,
clothoid arcs and Pythagorean hodographs can be found in Tsour-
dos et al. (2010). Compared to clothoids, Fermat's spiral can be used
in order to generate curvature-continuous paths with a very low
computational cost. The approach was further extended in Lekkas
et al. (2013) where an alternative parameterization of Fermat's
spiral was proposed, which are suitable for path planning.
 Category 3: Piecewise polynomial and spline paths.
The third category of methodologies pertains to generate tra-
jectories using piecewise polynomial and Splines. piecewise
Comments
Not possible for smooth transition.
Difficult transition maneuvers between straight and circular segments.
Linearly varying curvature over the path length, increased computational
cost.
Guaranted to satisfy the curvature constraints with increased computational
cost.
Curvature-continuous paths with a very low computational cost compared to
clothoids.
Twice differentiable, and the corresponding steering controls are piecewise
continuous. but not be able to provide complex curvature profile for long-
range global path planning.
Curvature continuous paths pass through the waypoints, but relocating one
waypoint induces changes along the path.
Similar to natural spline with the possibility of assigning the derivative values
at the control points. It does not have a continuous second derivative.
Continuous curvature without passing through all the control points, relo-
cating one waypoint only locally changes the path.
 Z. Zeng et al. / Ocean Engineering 110 (2015) 303–313
Table 5
Literature review of optimization techniques.
Optimization techniques Reference
Graph search methods Dijkstra Dijkstra, 1959
A* Carroll et al., 1992
Field D* Ferguson and Stentz, 2006
FM & LSM Pêtrès et al., 2007; Lolla et al.,
2012
APF Warren, 1990; Kruger et al.,
2007
RRT Rao and Williams, 2009
Evolutionary GA Alvarez et al., 2004;
Approaches PSO Witt and Dunbabin, 2008
QPSO Fu et al., 2012
polynomial and Spline curves need only a few variables (coordi-
nates of the control points) in order to define complicated curved
paths, consequently, both for path optimization and for on-line
implementation, it is easier to deal with small number of para-
meters when generating paths with minimum computational cost.
A family of sixth-order piecewise-constant polynomials is applied
to generate feasible trajectories in Qu et al. (2004). The resulting
trajectory is twice differentiable, and the corresponding steering
controls are piecewise continuous. However, this technique is only
tested for local analytical motion planning of car-like mobile robot, it
may not be able to provide complex curvature profile that is desired
for long-range global path planning. Natural splines pass through the
waypoints and also produce curvature continuous paths, however,
they do not possess local control which refers to the case where
relocating one waypoint induces changes throughout a larger part of
the path. In addition, the resulting paths are not very practical
(Fossen, 2011). An alternative is the Cubic Hermite Spline which
passes through all the waypoints and allows the derivative values at
the control points, and also permits local control over the path
(Wagner et al., 2010). The disadvantage of this Cubic Hermite Spline
is that it does not have a continuous second derivative.
B-splines can give paths a desired second derivative continuity
without passing through all the control points used to define it (Jolly
et al., 2009). The authors in (Lapierre and Jouvencel, 2008) presented
an implicit time-parameterization of the trajectory using a B-spline
representation. Designing an obstacle-avoiding B-spline path was
dealt with by (Antonelli et al., 2007), whereas the real-time mod-
ifications of a spline path is proposed in (Breivik and Fossen, 2007).
Another valuable characteristic of the adopted B-Spline curves
is that the curve is tangential to the control polygon at the starting
and ending points. This characteristic can be used in order to
define the starting direction of the curve, by inserting an extra
fixed point after the starting one. These two points can define the
direction of the curve at the corresponding region. This is essential
for the path planning of autonomous vehicles, as their traveling
angles are continuously defined. Consequently the direction of the
designed path line in the starting position must coincide with the
current direction of travel in this position, in order to ensure
curvature continuity of the whole path line.
5. Optimization techniques for path planning
Path planning in the robotics literature covers the topic of
finding a feasible and usually optimal path to allow a robot to move
autonomously from one location to another in the environment
(LaValle and Kuffner, 2001). This section presents a detailed litera-
ture review of the state-of-the-art AUV path planning techniques
with discussion of their assumptions and drawbacks. A brief
309
Completeness Comments
Resolution completeness Grid-search based scheme, discrete state transitions
Computationally expensive in high-dimensional problems
Resolution completeness Only allow linear cost function to preserve computational efficiency
Probabilistic resolution Fast but susceptible to local minima
completeness
Probabilistic completeness Fast and effective with high-dimensional configuration spaces, but
solutions are sub-optimal and often require further refinement.
Probabilistic completeness Practical in in high-dimensional problems, but may converge to a
suboptimal solution within a finite time
comparison of path planning techniques for AUV is available in
Table 5. Two important properties of path planning algorithms are
the completeness and the optimality of the algorithm. Two forms of
completeness are probabilistic completeness and resolution com-
pleteness. An algorithm is called Resolution completeness if it is
guaranteed to find an existing solution in finite time as long as the
resolution of an underlying grid is fine enough. Most resolution
complete planners are graph search methods such as Dijkstra, A*
and Field D*. In contrast, an algorithm is considered probabilistically
complete if the probability of finding a path approaches 100%.
Several sample-based methods, such as RRT and evolutionary
algorithms are probabilistically complete. The performance of a
probabilistically complete planner is shown by the rate of con-
vergence. Optimality is the property that the planner computes the
optimal path with respect to some criterion, e.g., minimal time,
energy consumption or distance. Probabilistic optimality and reso-
lution optimality are similarly to the definition of probabilistic
completeness and resolution completeness.
5.1. Graph search shemes
Graph-based methods are a classical path planning approach
that lies in the category of discrete optimal planning (Tovar et al.,
2007). A grid-shape graph represents the search space with the
edges labeled indicating the cost of traveling from a vertex to one
of its neighbors. Dijkstra's algorithm is probably the first graph
method adapted to search for a minimum cost paths, it computes
every possible path from a starting point to a specified destination
point (Dijkstra, 1959). With its heuristic searching ability, the A*
algorithm (Hart et al., 1968) has proven to be more efficient. The
heuristic function provides an estimate of the cost of the best
route that passes through a particular node. The algorithm keeps
track of the cost of the route leading up to a particular node along
with the heuristic cost function to determine which node it must
visit next.; Carroll et al. (1992) applied A* on a quad-tree search
space, which was adapted to the ocean currents field, i.e. it has
higher resolution where the ocean currents vary more spatially; or
more formally, where the gradient of the ocean currents is greater.
Overall, these grid-based graph search method are commonly
criticized for their discrete state transitions which unnaturally
constrain the motion of a vehicle to limited directions. There exists
a number of variants of A* that are worth mention. Any-angle
methods, like Theta* (Nash et al., 2007; Daniel et al., 2010), try to
obtain shorter paths alleviating the angle discretization problem
caused by the search grid. The Field D* algorithm uses a linear
interpolation-based method to allow continuous heading direc-
tions, but these variants of A* still not fix the problem of compu-
tationally expensive to employ in high-dimensional problems
(Ferguson and Stentz, 2006).
310 Z. Zeng et al. / Ocean Engineering 110 (2015) 303–313
5.2. Fast Marching and Level Set Methods (FM & LSM)
The FM algorithm can be regarded as a continuous version of
Dijkstra's algorithm. It uses a first order numerical approximation of
the nonlinear Eikonal equation. FM algorithm have been recently
applied for AUV path planning by Pêtrès et al. (2005), (2007). A
heuristically guided version of FM, known as FM*, maintains the
accuracy of the FM algorithm along with the efficiency of the A*
algorithm; however it is limited in that it uses a linear anisotropic
cost function to improve the algorithm computational efficiency.
The FM* scheme is improved in Soulignac et al. (2008), (2009),
Soulignac (2011) by using wavefront expansion to calculate shortest
time paths and also determines the departure time of the vehicle
from the starting point. The LSM is a more general technique than
the Fast Marching algorithm for wavefront expansion (Sethian,
2001). This method had been applied for path planning in flow
fields. The time-optimal path is generated by solving a particle
tracking equation backward in time after it evolves a front from the
vehicle's start location until it reaches the goal (Lolla et al., 2012).
The level set method provides the ability to solve more complex
problems, but it takes longer computation time than Fast Marching.
5.3. Artificial Potential Field (APF)
An artificial potential field for global path planning based on a
linear energy cost-function was originally proposed by Warren
(Warren, 1990). Since then, it has been widely used by the robotics
community and many problem specific developments have been
made to this algorithm (Barraquand et al., 1992). The key idea of this
approach is to introduce an artificial potential field on the obstacles
that prevents vehicles from getting very close to them, thus, gen-
erating safe paths. Kruger et al. (2007) then replaced the single term
cost-function with one that that incorporates a mixture of various
linear terms, including energy, obstacle regions, distance, time and
excess speed. Potential fields have also been used for underwater
path planning in Witt and Dunbabin (2008) with a cost function
measuring the total drag experienced by the vehicle, total travel
time and any obstacles in the field. After generating a feasible set of
tracks, an optimization is performed on these tracks. This algorithm
has the advantage of being inexpensive, thus allowing for easy real-
time computations to adapt the vehicle path. However, it has the
drawback of producing locally optimal solutions. Another problem
with potential field methods is their adaptation to dynamic ocean
currents. It is very inefficient to re-compute the potential field for
the whole map for each time instant.
5.4. Rapidly-exploring Random Trees (RRT)
Rapidly-exploring Random Trees (RRT) have also been used to
solve the path planning problem. RRT incrementally grow a tree to
explore the space until the tree branch reaches the goal location
(Kuffner and La Valle, 2000). A number of extensions like the
Dynamic RRT (Ferguson and Stentz, 2006) and Multi-particle RRT
(Zucker et al., 2007), have been made to facilitate robot navigation
in dynamic and uncertain environments. RRT has been applied to
both AUVs (Tan et al., 2005) and gliders (Rao and Williams, 2009)
path planning in dynamic flow-fields. It is interesting to note that
the RRT growth is inherently biased in the direction of ocean
currents. A comparison of RRT and grid-based method is also
presented in (Rao and Williams, 2009). The RRT algorithm is very
fast and effective to produce collision free paths to problems with
high-dimensional configuration spaces, but it does not assure
optimality and the paths found often requires further refinement.
5.5. Evolutionary algorithms
Evolutionary algorithms are another technique used for path
planning. The genetic algorithm (GA) (Nikolos et al., 2003; Alvarez
et al., 2004) and the particle swarm optimization (PSO) (Besada-Portas
et al., 2013; Roberge et al., 2014) are two well-known forms of evo-
lutionary algorithms that are generally recognized to be effective
optimization techniques for solving path planning problems. GA and
PSO algorithms are similar in the sense that they are population-based
search scheme and that they all depend on information sharing
among their population members to enhance their search processes
using a combination of deterministic and probabilistic rules. However,
PSO has its salient characteristics different from GA. PSO is a stochastic
evolutionary algorithm that does not incorporate survival of the fittest,
and there is no conventional evolutionary operators such as crossover
and mutation. For PSO, all particles are retained through the course of
the run and each particle adjusts its searching in the space in terms of
its own searching experience and its companions' searching experi-
ence. A detailed comparison between PSO and GA can be found in
(Eberhart and Shi, 1998). Quantum-behaved particle swarm optimi-
zation (QPSO) is recognized as an improved version of the original
PSO. It differs in that QPSO assumes that every particle in the swarm
has quantum behavior instead of using the conventional position and
velocity update rules employed in PSO. Fu et al. (2012) applied the
QPSO for path planning and showed that it has superior performance
compared to the standard PSO and GA algorithms. However, Fu's work
is focus on UAV-oriented applications, it do not consider ocean cur-
rents information, nor the effects of currents on the AUV completing
the mission.
Planning for AUVs operating in a large geographical area is a
typical large-scale optimization problem. Evolutionary computation
techniques have been proven to be an effective way of dealing with
NP-hard problem. Their computational cost grows linearly with the
number of vehicles and geometrically with spatial dimensions. Also,
they can escape from local minima. Evolutionary algorithms are
population based optimization techniques, they search for a solu-
tion in parallel, which allows them to be implemented on parallel
machines to achieve super linear speed-up with the number of
processors (Roberge et al., 2014). Their drawback is that they may
converge to a suboptimal solution within a finite time.
6. Open problems and research challenges
In the last decade, significant progress has been made to
improve the capability of the AUVs to have longer mission dura-
tions as well as higher levels of autonomy. This survey has
reviewed the current state of the art in the AUV path planning
problem and provided a detailed overview of published papers,
with a particular focus on shape and properties of the path that
have been implemented and optimization techniques for path
planning. The key conclusions of this paper are the following:
 As illustrated from this review, although several path planning
and re-planning methods have been previously proposed for
autonomous vehicles, several difficulties still remain for AUV-
oriented applications. Path planning and re-planning for AUVs
that operate across a large geographical area is a large-scale
optimization problem. The computational requirements grow
exponentially for solving high dimensional problems. In order
to speed up the planning process and reduce the memory
requirement, most conventional path planning approaches
project the 3D environment to 2D space (Alvarez et al., 2004;
Smith et al., 2010, 2013) However, this 2D space cannot
completely embody all the 3D information of the ocean envir-
onment. The specific characteristics of ocean current
 Z. Zeng et al. / Ocean Engineering 110 (2015) 303–313
environments and various maneuverability of different vehicles
pose more challenges to cooperative path planning of multiple
AUVs. Evolutionary algorithms have been proven to be an
efficient and effective way of dealing with non-deterministic
polynomial-time (NP) hard problems (Besada-Portas et al.,
2013). Also, evolutionary algorithms are population based
optimization techniques and amenable to be implemented on
a parallel machine to achieve super linear speed-up with the
number of processors (Roberge et al., 2014). However, they have
the drawbacks of inconsistency and incompleteness of search-
ing for the optimal path. With acknowledge of this limitation of
the evolutionary path planner, therefore methodologies should
be researched to improve evolutionary planners with increased
path searching efficiency, and to have better robustness and
quality.
 It was difficult to compare the reviewed optimization algo-
rithms for path planning. More work needs to take place on the
development of benchmarks to compare algorithms and to
provide some design standards for path planning technologies.
 Many papers have been published in the AUV path planning
areas, but very few of them have reported convincing field trail
results of AUVs operating in dynamic, cluttered, and uncertain
ocean environments. It is hard to tell that the technologies
developed so far are reliable enough to deal with complex ocean
environments or only capable of dealing particular missions.
 There is a gap between off-line and on-line path planning.
Indeed, a significant amount of work has been done in the
design of off-line path planning techniques AUVs. However, most
of the on-line path planning systems only have been validated in
simulations, or very basic lake test without convincing results.
We believe that significant research is also needed to develop
and implement the on-line planning systems.
 To data, the literature on cooperative use of multiple vehicles
mainly focus on control, localization and navigation, there is
still considerable work to be done to develop advanced methods
for cooperative multiple AUVs path planning, and explicitly
addressing the problem of simultaneous rendezvous is timely
and challenging.
6.1. Path planning in complex ocean environments
Most current research activities focus on path planning of AUVs
operating in static or quasi-static environments. Although there
have been some interesting results on AUV autonomous naviga-
tion through variable currents, the current path planning and
guidance technologies have not yet reached the level of robustness
and reliability required for real-world applications. There is thus a
real need to developing and evaluating efficient but rigorous path
planners that generate optimal paths for AUVs in realistic ocean
environments. The challenge here is to develop computationally
efficient and rigorous frameworks that accommodate both the
environmental constraints and vehicle manuverbility while at the
same time providing an optimal path for the vehicle (Zeng et al.,
2014a, 2014b, 2014c). The ocean environments are composed of
strong currents field, irregularly shaped terrains and obstacles, the
position of which may be dynamic and uncertain.
6.2. On-line path replanning in spatiotemporal environments
Despite active research on AUV path planning and guidance, most
current work still use off-line planning system. From the reviewed
literature, we found that the majority of existing algorithms do not
include much results and discussion of on-line implementation. The
main problem of implementing and demonstrating effective on-line
path planning systems may be optimization algorithms that are
complete, optimal, and that consider vehicle dynamics are
311
computationally expensive and their fast reaction to the dynamic
environments is very challenging (Zeng et al., 2014a, 2014b, 2014c).
Therefore, there is still much work to be done on developing advanced
methods that enable an AUV to adapt and regenerate its trajectory
during the course of the mission using continuously updated current
profiles from on-board sensors, such as a Horizontal Acoustic Doppler
Velocity Logger (Garau et al., 2006). Again, the ocean environments
may include spatiotemporal currents, irregularly shaped terrains as
well as obstacles whose position coordinates are uncertain. AUV may
instructed to rendezvous with a moving target, such as a mother ship
or autonomous surface vehicle.
6.3. Path planning for multi-AUVs rendezvous
Early research on path planning of a single AUV operating in
simple environments. In recent years, there have been fewer pub-
lished papers on the planning of a single AUV, in fact, many
researchers have shifted their research interest to the efficient and
robust path planning algorithms, mission schedule systems, and
higher-level guidance modules for multiple AUVs. We recommend
more efforts in rigorously developing an effective path planning
guidance system for multiple marine vehicles rendezvous in ocean
environments (Zeng et al., 2014a, 2014b, 2014c). This system should
be able to generate trajectories for multiple vehicles with minimal
time usage over all participating vehicles and simultaneous arrival
of the vehicles at their selected rendezvous destination. This system
also need to be lightweight in terms of running time and capable of
incorporating different factors influencing a given mission, such as
vehicles' dynamic constraints and environmental conditions.
This paper reviews the current and future needs of persistent
presence of AUVs. The paper highlights challenges that need to be
addressed and suggested the new autonomous planning techni-
ques intended to facilitate the capability of the AUVs to have
longer mission durations as well as higher levels of autonomy. The
next stage in this work is to apply one or multiple benchmark
problems to the reviewed path planning algorithms, and analyse
the weaknesses and strengths of each technique to determine the
most appropriate algorithm for AUV path planning.
Acknowledgment
This study is supported by the National Natural Science Foun-
dation of China (NSFC) (No. 51279107) and the Research Fund for
Science and Technology Commission of Shanghai Municipality
(STCSM) (No. 13dz1204600).
References
Autonomous underwater vehicle-MUNIN AUV. 〈http://www.km.kongsberg.com/ks/
web/nokbg0240.nsf/AllWeb/48D837005CB784DFC1257C970041051F?
OpenDocument〉.
Abu Sharkh, S.M., Griffiths, G., 2002. Energy storage systems for unmanned
underwater vehicles. Underwater Technol. 25 (3), 143–148.
Alvarez, A., Caiti, A., Onken, R., 2004. Evolutionary path planning for autonomous
underwater vehicles in a variable ocean. IEEE J. Ocean. Eng. 29 (2), 418–429.
Antonelli, G., Chiaverini, S., Fusco, G., 2007. A fuzzy-logic-based approach for
mobile robot path tracking. IEEE Trans. Fuzzy Syst. 15 (2), 211–221.
Bluefin-9 Product Sheet-Bluefin Robotics.〈http://www.bluefinrobotics.com/assets/
Downloads/Bluefin-9-Product-Sheet.pdf〉.
Bluefin-21 Product Sheet-Bluefin Robotics.〈http://www.bluefinrobotics.com/pro
ducts/bluefin-21/〉.
Bakolas, E. and P.Tsiotras (2010). Time-optimal synthesis for the Zermelo–Markov–
Dubins problem: The constant wind case. In: Proceedings of the 2010 American
Control Conference, ACC2010.
Bakolas, E., Tsiotras, P., 2013. Optimal partitioning for spatiotemporal coverage in a
drift field. Automatica 49 (7), 2064–2073.
Barraquand, J., Langlois, B., Latombe, J.-C., 1992. Numerical potential field techni-
ques for robot path planning. IEEE Trans. Syst., Man Cybern. 22 (2), 224–241.
312 Z. Zeng et al. / Ocean Engineering 110 (2015) 303–313
Besada-Portas, E., De La Torre, L., Moreno, A., Risco-Martín, J.L., 2013. On the per-
formance comparison of multi-objective evolutionary UAV path planners. Inf.
Sci. 238, 111–125.
Breivik, M. and T. I.Fossen (2007). Applying missile guidance concepts to motion
control of marine craft. IFAC Proceedings Volumes(IFAC-PapersOnline).
Bruyninckx, H. and D.Reynaerts (1997). Path planning for mobile and hyper-
redundant robots using Pythagorean hodograph curves. In: Proceedings of
the International Conference on Advanced Robotics, ICAR.
Carroll, K. P., S. R.McClaran, E. L.Nelson, D. M.Barnett, D. K.Friesen and G.William
(1992). AUV path planning: an A* approach to path planning with considera-
tion of variable vehicle speeds and multiple, overlapping, time-dependent
exclusion zones. In: Proceedings of the 1992 Symposium on Autonomous
Underwater Vehicle Technology, 1992. AUV'92., IEEE.
Chitsaz, H. and S. M.LaValle (2007). Time-optimal paths for a dubins airplane. In:
Proceedings of the IEEE Conference on Decision and Control.
Cowlagi, R. V. and P.Tsiotras (2008). Multiresolution path planning with wavelets: A
local replanning approach. In: Proceedings of the American Control Conference.
Dahl, A. (2013). Path Planning and Guidance for Marine Surface Vessels.
Daniel, K., Nash, A., Koenig, S., Felner, A., 2010. Theta*: any-angle path planning on
grids. J. Artif. Intell. Res. 39, 533–579.
Dijkstra, E.W., 1959. A note on two problems in connexion with graphs. Numer.
Math. 1 (1), 269–271.
Dubins, L.E., 1957. On curves of minimal length with a constraint on average cur-
vature, and with prescribed initial and terminal positions and tangents. Am. J.
Math. 79 (3), 497–516.
Eberhart, R.C., Shi, Y., 1998. Comparison between genetic algorithms and particle
swarm optimization. Springer, Berlin, Heidelberg, In Evolutionary Program-
ming VII (pp. 611–616).
Fallon, M.F., Papadopoulos, G., Leonard, J.J., Patrikalakis, N.M., 2010. Cooperative
AUV navigation using a single maneuvering surface craft. Int. J. Robot. Res. 29
(12), 1461–1474.
Ferguson, D., N.Kalra and A.Stentz (2006). Replanning with RRTs. In: Proceedings
2006 IEEE International Conference on Robotics and Automation, 2006.
ICRA2006.
Ferguson, D. and A.Stentz (2006). Anytime RRTs. In: Proceedings of the IEEE
International Conference on Intelligent Robots and Systems.
Ferguson, D., Stentz, A., 2006. Using interpolation to improve path planning the
field D* algorithm. J. Field Robot. 23 (2), 79–101.
Fiorelli, E., Leonard, N.E., Bhatta, P., Paley, D.A., Bachmayer, R., Fratantoni, D.M.,
2006. Multi-AUV control and adaptive sampling in Monterey Bay. IEEE J. Ocean.
Eng. 31 (4), 935–948.
Fossen, T. I. (2011). Handbook of Marine Craft Hydrodynamics and Motion Control.
Fraichard, T., Scheuer, A., 2004. From Reeds and Shepp's to continuous-curvature
paths. IEEE Trans. Robot. 20 (6), 1025–1035.
Fu, Y., Ding, M., Zhou, C., 2012. Phase angle-encoded and quantum-behaved particle
swarm optimization applied to three-dimensional route planning for UAV. IEEE
Trans. Syst., Man Cybern. Part A: Syst. Humans 42 (2), 511–526.
M.E. Furlong D. Paxton P. Stevenson M. Pebody S.D. McPhail J. Perrett. Autosub Long
Range: A long range deep diving AUV for ocean monitoring 2012 IEEE/OES
Autonomous Underwater Vehicles 2012. AUV 2012.
Garau, B., A.Alvarez and G.Oliver (2006). AUV navigation through turbulent ocean
environments supported by onboard H-ADCP. In: Proceedings of the IEEE
International Conference on Robotics and Automation, Orlando, FL.
Garau, B., Bonet, M., Alvarez, A., Ruiz, S., Pascual, A., 2009. Path planning for
autdndmdus underwater vehicles in realistic oceanic current fields: application
to gliders in the Western Mediterranean sea. J. Mar. Res. 6 (2), 5–21.
Griffiths, G., McPhail, S.D., 2011. AUVs for depth and distance: Autosub6000 and
autosub long range. Sea Technol. 52 (12), 27–30.
P.E. Hagen N. Størkersen K. Vestgård P. Kartvedt The HUGIN 1000 autonomous
underwater vehicle for military applications Ocean. In: Proceedings of the
Conference Record (IEEE) 2003.
Harary, G., Tal, A., 2012. 3D Euler spirals for 3D curve completion. Comput. Geo.:
Theor. Appl. 45 (3), 115–126.
Hart, P.E., Nilsson, N.J., Raphael, B., 1968. A formal basis for the heuristic determi-
nation of minimum cost paths. Syst. Sci. Cybern., IEEE Trans. 4 (2), 100–107.
Hasvold, Ø., Størkersen, N.J., Forseth, S., Lian, T., 2006. Power sources for autono-
mous underwater vehicles. J. Power Sour. 162, 935–942 2 SPEC. ISS..
B.W. Hobson J.G. Bellingham B. Kieft R. McEwen M. Godin Y. Zhang Tethys-class
long range AUVs-extending the endurance of propeller-driven cruising AUVs
from days to weeks 2012 IEEE/OES Autonomous Underwater Vehicles 2012.AUV
2012.
Hota, S., Ghose, D., 2013. Optimal trajectory generation for convergence to a rec-
tilinear path. J. Intell. Robot. Syst.: Theor. Appl. 75 (2), 223–242.
Incze, M.L., 2011. Lightweight autonomous underwater vehicles (AUVs) performing
coastal survey operations in REP 10A. Ocean Dyn. 61 (11), 1955–1965.
Jolly, K.G., Sreerama Kumar, R., Vijayakumar, R., 2009. A Bezier curve based path
planning in a multi-agent robot soccer system without violating the accelera-
tion limits. Robot. Autonom. Syst. 57 (1), 23–33.
Joung, T., K.Sammut, F.He and S.K.Lee (2009). A study on the design optimization of
an AUV by using computational fluid dynamic analysis. In: Proceedings of the
International Offshore and Polar Engineering Conference.
Joung, T.H., Sammut, K., He, F., Lee, S.K., 2012. Shape optimization of an autonomous
underwater vehicle with a ducted propeller using computational fluid
dynamics analysis. Int. J. Naval Architect. Ocean Eng. 4 (1), 44–56.
Jung, D. and P.Tsiotras (2008). Multiresolution on-line path planning for small
unmanned aerial vehicles. In: Proceedings of the American Control Conference.
Kanayama, Y., Yuta, Si, 1988. Vehicle path specification by a sequence of straight
lines. IEEE J. Robot. Autom. 4 (3), 265–276.
Knudson, M., Tumer, K., 2011. Adaptive navigation for autonomous robots. Robot.
Autonom. Syst. 59 (6), 410–420.
T.B. Koay M. Chitre Energy-efficient path planning for fully propelled AUVs in
congested coastal waters OCEANS 2013 MTS/IEEE Bergen: The Challenges of the
Northern Dimension. 2013.
Kruger, D., R.Stolkin, A.Blum and J.Briganti (2007). estuarine environments. In:
Proceedings of the -IEEE International Conference on Robotics and Automation.
KuffnerJr, J. J. and S. M.La Valle (2000). RRT-connect: an efficient approach to single-
query path planning. In: Proceedings of the -IEEE International Conference on
Robotics and Automation.
Kuwata, Y., T.Schouwenaars, A.Richards and J.How (2005). Robust constrained
receding horizon control for trajectory planning. Collection of Technical Papers-
AIAA Guidance, Navigation and Control Conference.
Kwak, S. H., Ong, S.M., McGhee, R.B., 1990. A mission planning expert system for an
autonomous underwater vehicle. In: Proceedings of the Symposium on -
Autonomous Underwater Vehicle Technology, 1990. AUV '90.
Lapierre, L., Jouvencel, B., 2008. Robust nonlinear path-following control of an AUV.
IEEE J. Ocean. Eng. 33 (2), 89–102.
LaValle, S.M., Kuffner Jr, J.J., 2001. Randomized kinodynamic planning. Int. J. Robot.
Res. 20 (5), 378–400.
Lekkas, A. M., A. R.Dahl, M.Breivik and T. I.Fossen (2013). Continuous-Curvature
Path Generation Using Fermat.
Lekkas, A.M., Fossen, T.I., 2014. Integral LOS path following for curved paths based
on a monotone cubic hermite spline parametrization. IEEE Trans. Control Syst.
Technol. 22 (6), 2287–2301.
Lolla, T., M.P. Ueckermann, K.Yi|git, P. J.HaleyJr and P.F. J.Lermusiaux (2012). Path
planning in time dependent flow fields using level set methods. In: Proceedings
of the -IEEE International Conference on Robotics and Automation.
Navy'Mine-Hunter' AUV Sets Mission Endurance Record. 〈http://www.nrl.navy.mil/
media/news-releases/2013/navy-mine-hunter-auv-sets-mission-endurance-
record〉.
A. Nash K. Daniel S. Koenig A. Feiner Theta*: Any-angle path planning on grids. In:
Proceedings of the National Conference on Artificial Intelligence. 2007.
Nikolos, I.K., Valavanis, K.P., Tsourveloudis, N.C., Kostaras, A.N., 2003. Evolutionary
algorithm based offline/online path planner for uav navigation. IEEE Trans Syst.,
Man Cybern., Part B: Cybern. 33 (6), 898–912.
Nygaard, K.H., 2014. Meet the family. Int. Ocean Syst. 18 (1), 6–7.
Pêtrès, C., Pailhas, Y., Patrón, P., Petillot, Y., Evans, J., Lane, D., 2007. Path planning for
autonomous underwater vehicles. IEEE Trans. Robot. 23 (2), 331–341.
C. Pêtrès Y. Pailhas Y. Petillot D. Lane Underwater path planing using fast marching
algorithms Oceans 2005-Europe 2005.
Pereira, A.A., Binney, J., Hollinger, G.A., Sukhatme, G.S., 2013. Risk-aware path
planning for autonomous underwater vehicles using predictive ocean models. J.
Field Robot. 30 (5), 741–762.
Pereira, A. A., J.Binney, B. H.Jones, M.Ragan and G. S.Sukhatme (2011). Toward risk
aware mission planning for autonomous underwater vehicles. IEEE Interna-
tional Conference on Intelligent Robots and Systems.
Qu, Z., Wang, J., Plaisted, C.E., 2004. A new analytical solution to mobile robot
trajectory generation in the presence of moving obstacles. IEEE Trans. Robot. 20
(6), 978–993.
Rao, D. and S. B.Williams (2009). Large-scale path planning for underwater gliders
in ocean currents. Proceedings of the 2009 Australasian Conference on Robotics
and Automation, ACRA2009.
Reeds, J., Shepp, L., 1990. Optimal paths for a car that goes both forwards and
backwards. Pacific Journal of Mathematics 145 (2), 367–393.
Roberge, V., Tarbouchi, M., Allaire, F., 2014. Parallel hybrid metaheuristic on shared
memory system for real-time UAV path planning. Int. J. Comput. Intell. Appl. 13
(2).
Sethian, J.A., 2001. Evolution, implementation, and application of level set and fast
marching methods for advancing fronts. J. Comput. Phys. 169 (2), 503–555.
Smith, R.N., Chao, Y., Li, P.P., Caron, D.A., Jones, B.H., Sukhatme, G.S., 2010. Planning
and implementing trajectories for autonomous underwater vehicles to track
evolving ocean processes based on predictions from a regional ocean model.
Int. J. Robot. Res. 29 (12), 1475–1497.
Smith, R.N., Kelly, J., Nazarzadeh, K., Sukhatme, G.S., 2013. An investigation on the
accuracy of Regional Ocean Models through field trials. In: Proceedings of the
IEEE International Conference on Robotics and Automation.
Soulignac, M., 2011. Feasible and optimal path planning in strong current fields.
IEEE Trans. Robot. 27 (1), 89–98.
Soulignac, M., P.Taillibert and M.Rueher (2008). Adapting the wavefront expansion
in presence of strong currents. In: Proceedings of the -IEEE International
Conference on Robotics and Automation.
Soulignac, M., P.Taillibert and M.Rueher (2009). Time-minimal path planning in
dynamic current fields. In: Proceedings of the -IEEE International Conference
on Robotics and Automation.
Sprunk, C., 2008. Planning motion trajectories for mobile robots using splines.
University of Freiburg.
Stokey, R. P., A.Roup, C.Von Alt, B.Allen, N.Forrester, T.Austin, R.Goldsborough, M.
Purcell, F.Jaffre, G.Packard and A.Kukulya (2005). Development of the REMUS 600
autonomous underwater vehicle. In: Proceedings of MTS/IEEE OCEANS, 2005.
Teledyne Gavia. 〈http://www.teledynegavia.com/index.php/product_dashboard/auvs〉.
Tan, C. S., R.Sutton and J.Chudley (2005). Quasi-random, Manoeuvre-based motion
planning algorithm for autonomous underwater vehicles. In: IFAC Proceedings
Volumes (IFAC-PapersOnline).
 Z. Zeng et al. / Ocean Engineering 110 (2015) 303–313
Taylor, M. and A.Wilby (2011). Design considerations and operational advantages of
a modular AUV with synthetic aperture sonar. OCEANS'11-MTS/IEEE Kona,
Program Book.
Techy, L., Woolsey, C.A., 2009. Minimum-time path planning for unmanned aerial
vehicles in steady uniform winds. J. Guidance, Control, Dyn. 32 (6), 1736–1746.
Thompson, D. R., S.Chien, Y.Chao, P.Li, B.Cahill, J.Levin, O.Schofield, A.Balasuriya, S.
Petillo, M.Arrott and M.Meisinger (2010). Spatiotemporal path planning in
strong, dynamic, uncertain currents. In: Proceedings of the -IEEE International
Conference on Robotics and Automation.
Tovar, B., Murrieta-Cid, R., LaValle, S.M., 2007. Distance-optimal navigation in an
unknown environment without sensing distances. IEEE Trans. Robot. 23 (3),
506–518.
Tsourdos, A., White, B., Shanmugavel, M., 2010a. Cooperative path planning of
unmanned aerial vehicles. John Wiley & Sons, Vol. 32.
Wagner, P., J.Kordas, V.Michna and J.Kotzian (2010). Motion control for robots based
on cubic hermite splines in real-time. In: IFAC Proceedings Volumes (IFAC-
PapersOnline).
C.W. Warren A technique for autonomous underwater vehicle route planning
Proceedings of the Symposium on Autonomous Underwater Vehicle Technol-
ogy 1990.
Westwood, J. (2010). "The AUV market place."Presentation at Oceanology Interna-
tional, London, 10th March.
313
Witt, J. and M.Dunbabin (2008). Go with the flow: Optimal AUV path planning in
coastal environments. In: Proceedings of the 2008 Australasian Conference on
Robotics and Automation, ACRA2008.
Wzorek, M. and P.Doherty (2006). Reconfigurable path planning for an autonomous
unmanned aerial vehicle. In: Proceedings of the -2006 International Conference
on Hybrid Information Technology, ICHIT2006.
Wzorek, M., J.Kvarnström and P.Doherty (2010). Choosing path replanning strate-
gies for unmanned aircraft systems. In: ICAPS 2010-Proceedings of the 20th
International Conference on Automated Planning and Scheduling.
Zeng, Z., Lammas, A., Sammut, K., He, F., Tang, Y., 2014a. Shell space decomposition
based path planning for AUVs operating in a variable environment. Ocean Eng.
91 (0), 181–195.
Zeng, Z., A.Lammas, K.Sammut, F.He, Y.Tang and Q.Ji (2014b). Path planning for
rendezvous of multiple AUVs operating in a variable ocean. In: Proceedings of
the 2014 IEEE 4th Annual International Conference on Cyber Technology in
Automation, Control and Intelligent Systems (CYBER), IEEE.
Zeng, Z., Sammut, K., Lammas, A., He, F., Tang, Y., 2014c. Efficient Path Re-planning
for AUVs Operating in Spatiotemporal Currents. J. Intell. Robot. Syst. 79 (1),
135–153.
Zucker, M., J.Kuffner and M.Branicky (2007). Multipartite RRTs for rapid replanning
in dynamic environments. In: Proceedings of the -IEEE International Con-
ference on Robotics and Automation.