See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/224305662

The Development and Ocean Testing of an AUV Docking Station for a 21 AUV

Conference Paper · November 2007

DOI: 10.1109/OCEANS.2007.4449318 · Source: IEEE Xplore

CITATIONS READS
39 2,460

7 authors, including:

B. W. Hobson Jon A Erickson

Monterey Bay Aquarium Research Institute Eli Lilly
46 PUBLICATIONS   627 CITATIONS    34 PUBLICATIONS   1,367 CITATIONS   

SEE PROFILE SEE PROFILE

Thomas Hoover Lance Ryan Mcbride

King Abdullah University of Science and Technology G.E.
7 PUBLICATIONS   125 CITATIONS    12 PUBLICATIONS   191 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

A New AAAS Robotics Journal View project

Tethys Long Range AUV View project

All content following this page was uploaded by B. W. Hobson on 29 September 2014.

The user has requested enhancement of the downloaded file.

 The Development and Ocean Testing of an AUV
Docking Station for a 21” AUV
Brett W. Hobson, Robert S. McEwen, Jon Erickson, Thomas Hoover,
Lance McBride, Farley Shane and James G. Bellingham
Monterey Bay Aquarium Research Institute
7700 Sandholdt Road, Moss Landing, CA 95039-9644
[email protected], [email protected], [email protected]

Abstract-The Monterey Bay Aquarium Research Institute

(MBARI) has developed an AUV docking station for a 21-inch
(54 cm) diameter AUV. The system was designed for operation
with cabled undersea observatories in water depths up to 4km
deep and has been demonstrated in the open ocean, though at
much shallower depths. The program demonstrated successful
autonomous homing and docking, data downloads, uploading of
new mission plans, battery recharging, and complete power
cycling of the AUV. We describe the design, and at-sea tests.
I. INTRODUCTION
Autonomous Underwater Vehicles (AUVs) are playing a
growing role in the ocean sciences. At MBARI, a 6 km depth-
rated mapping AUV is routinely generating bathymetry maps
with unprecedented precision, resolution and speed, and a
second AUV performs weekly missions to measure the
Figure 1: AUV in the dock during testing in MBARI's 1.4 million liter
physical, chemical, and biological measurements of the upper seawater test tank
water column in Monterey Bay. Both of these AUVs are
characteristic of a class of vehicles carrying sophisticated, but System (MARS) cabled observatory and the Monterey
power intensive instrumentation. The high power draw of the Ocean Observing System (MOOS) mooring-based
instrumentation limits today’s AUVs to endurances of roughly observatory. MARS is a testbed for subsequent cabled
one day. This means that for all but near-shore operations, the observatories, and provides an early opportunity to develop
vehicles must be attended by a ship. As a consequence, the and demonstrate the AUV docking capability in a logistically
fundamental utility of the AUV is limited by ship support benign environment. The compatibility of MARS and the
requirements. In some cases, the issue is cost. In other cases, docking system should ensure that the system is well suited to
the need for ship support makes activities impossible, for being deployed on future Regional scale Cabled Observatories
example in high sea states, or when events cannot be predicted (RCOs). The MOOS mooring system is also capable of
in advance for scheduling ship time. Our docking station is supporting docking, and while MOOS is unable to supply as
designed to link to the power and communication much power to a docking station as a shore-cabled
infrastructure of a seafloor observatory, enabling the AUV to observatory, it has the advantage of being much more easily
operate independent of a surface vessel for extended periods. relocated and deployed in regions of interest.
The docking station allows the vehicle to connect to either a In the tests described below, the AUV was able to
mooring or a cabled instrumentation node. Once attached, the successfully dock and undock repeatedly in the open ocean,
AUV can recharge its batteries, download data, upload new without assistance from the surface. Shore operators
instructions, and safely moor until the next mission. Once its downloaded mission data and uploaded the next mission
batteries are recharged, the AUV can be deployed nearly through an underwater Ethernet link, charged the battery, and
instantaneously. Instead of being limited to studying static or completely powered the vehicle down and back up. Figure 1
repetitive subjects, the AUV can be quickly dispatched to shows the dock and vehicle in a test tank.
record episodic events such as canyon turbidity flows or The dock worked well despite the fact that it was designed
plankton blooms or other event response applications, greatly for a permanent deployment in a deeper more benign
increasing the scientific utility of an AUV. environment at the MARS cabled observatory at 900 m depth
The move toward seafloor observatories is international in in Monterey Bay.
scope, and creates the possibility that power and
communication infrastructure will be available on the seafloor
in a variety of interesting regions. At MBARI, the focus has
been the development of the Monterey Accelerated Research

0-933957-35-1 ©2007 MTS

 docking, requiring a communication system. This dock must
also transfer power and data across a moving interface, most
likely using a slip-ring.
Guided by the research interests of MBARI scientists and
the opportunities afforded by the MARS observatory we
focused on development of a dock for deep ocean
deployments. From previous experiences of developing,
deploying and maintaining complex equipment in the deep
ocean it was clear that the dock must be as reliable as possible.
The easiest way to ensure reliability was to design a dock that
was as simple as possible; therefore we chose the fixed cone
design.
III. THE AUV
The AUV used for our docking work is a Dorado/Bluefin 21”
(54 cm) diameter type, which has a Gertler body shape, Series
58 Model 4154, with a cylindrical midsection inserted at the
point of maximum diameter, and a ring-wing tail [9], [10].
Figure 2: Dock being recovered after 5 months in Monterey Bay The AUV docking configuration of this vehicle is 358 cm long,
and weighs 640 kg including entrained water. Power storage is
II. DOCKING SYSTEM DESIGN
provided by a single, 2 kW-hr pressure tolerant lithium
AUV docking stations are a multiplier for one of the polymer battery pack that provides a maximum range of
fundamental features of AUVs that make them so powerful for approximately 30 km at 1.5 m/s. The vehicle must maintain a
ocean sampling: long-term, un-attended operations. Therefore, minimum speed of approximately 0.8 m/s for controllability
people have been proposing and demonstrating AUV Docking and to counteract 3.5 kg of positive buoyancy. The ring-wing
for almost as long as AUVs have been operational. The propulsion unit, shown in Figure 3, generates 52 N of thrust at
Remus operations from a docking station deployed on the Leo 300 rpm using a fractional horsepower DC brushless motor
15 cabled observatory in the mid 1990s, described by Stokey with a 10:1 reduction gear. A 2 degree-of-freedom gimbal
in [1], demonstrated the feasibility and scientific utility. Since provides control in both vertical and horizontal directions. The
then, many designs have been proposed and several groups control system consists of outer and inner proportional-
have demonstrated basic functionality, but to the best of our integral-derivative (PID) loops for both heading and depth.
knowledge, nobody has conducted sustained operations of an The vehicle navigation system is essentially Doppler velocity
AUV from a docking station for science or industry
applications.
The designs put forth so far either hook to a cable or pole or
USBL Transceiver
fly the AUV into a tube or garage. Lambiotte [2] and
Fukasawa [3] demonstrated a hook capture mechanisms for
the Morpheus and Marine Bird respectively, and a nose hook Data transfer
was used for an Odyssey by Singh et al. at WHOI [4]. The antenna
Remus dock at Leo 15 used a cone to align the vehicle for
docking in a tube. Cowen also used a cone dock for the
Flying Plug [5]. A variation of the cone dock, the weather-
vaning cone described by Coulson in [6], employs a rotating
joint that allows the cone and tube to align itself with the
ambient current. Another variant of the cone dock is a
grappling system described in [7] and [8] that requires an
AUV capable of precise hovering.
We considered three classes of dock for MBARI’s torpedo-
like flying vehicle: the pole, the weather-vaning cone, and the
fixed cone. The pole is omni-directional, and eases the
constraint on depth control, but requires the vehicle to carry an
external, complex, motorized latching device. Establishing the Latch/Charging
power and data connection is difficult since the docked pin receptacle
vehicle is not rigidly attached to the pole. A weather-vaning
cone is also omni-directional, but securely captures the
vehicle, making power and data connections easier. However Figure 3: Solid model of Docking AUV with 10” long Docking Section
the dock must communicate its heading to the vehicle prior to

0-933957-35-1 ©2007 MTS

log (DVL)-aided dead reckoning, with periodic global
positioning system (GPS) updates, which require surfacing.
The expected accuracy is 4% of distance traveled (DT),
circular error probability (CEP), with the navigation and
control sampling rate of 5 Hz.
A. Homing System
The AUV homes to the dock using an Ultra-Short Base Line
(USBL) sonar transceiver mounted in the vehicle nose, shown
as the magenta cylinder in Figure 3. MBARI selected a 4 km
rated Sonardyne Scout USBL homing system, although other
types of homing systems have been proven, such as an
electromagnetic system [11], and optical systems [5] and [7].
The USBL has the advantage of much greater range, and is
also reliable, mature, and commercially available. The only
drawback is that the USBL selected is relatively large and
takes up valuable space in the nose of the AUV. The USBL
provides a bearing to its beacon in both the horizontal and
vertical planes with a specified accuracy of .28 degrees root
mean square (RMS). The maximum sampling frequency of the
USBL is 1 Hz, its maximum range was about 2000 m and the
field of view was ±85
degrees. Another factor in the selection
of the Sonardyne Scout is that MBARI has already placed and
surveyed over 100 acoustic beacons in the Monterey Bay that
are compatible with this system. In addition to AUV docking, Figure 4: Cross section of AUV Dock
the USBL is useful as a general navigation instrument; since it philosophy of keeping the dock as simple as possible. Even
provides position fixes to bound steadily growing error of with the latch/charge pin located in the dock, the entire
DVL aided dead-reckoning navigation, much like GPS, but docking system has only a single moving part with an
without surfacing or stopping. associated drive motor.
B. Data Transfer III. THE DOCK
The AUV transfers data to the dock using modified off-the-
shelf wireless Ethernet (802.11b) radios. This thru-water, The MBARI design approach keeps the dock as simple as
contact-less system uses one antenna that is embedded in the possible and puts the complexity on the vehicle, since it is
top of the vehicle and a mate correspondingly located in the easier to retrieve for maintenance or repair. The dock is a
top of the docking tube. Seawater tank tests conducted at benthic, fixed-heading cone design that provides secure
MBARI achieved a transfer rate of 10 Mb/second in seawater parking and a rigid data/power connection. The drawback to
with a 2.5 cm antenna separation. The system also worked this design is that the vehicle must fly a near-bottom precision
very well in the ocean trials. approach along a fixed flight path, possibly in a cross current,
and must therefore contain the necessary electronics and
C. Power Transfer control algorithms.
Power transfer and mechanical latching were both The body of the dock is a fiberglass tube of diameter 57 cm,
accomplished by having a latch/charge pin inserted into the 3 cm wider than the vehicle’s diameter. The dock’s entry is a
AUV from the dock. This system is a slightly modified compliant, open frame cone, built of fiberglass staves held in
version of the system developed by Florida Atlantic place by two tubular stainless-steel rings with an entrance
University (FAU) [7]. The system works just like many of the diameter of two meters. This entrance diameter was selected
popular electric toothbrush charging stations. When the latch based on observations that the vehicle was passing well within
pin is inserted into the AUV, an inductor coil that is embedded two meters of the transponder during initial homing tests [12],
in the tip of the pin is engaged between two mating inductor and the predicted position deviation of the vehicle based on
coils inside the docking payload section of the AUV. The simulations that used historical current data for the near-shore
system then transfers power inductively through the sea water area we planned to conduct tests. The cone and tube are
without a metal-to-metal connection. This system was mounted on a 2 m tall aluminum tripod (Fig.4). A pressure
demonstrated to transfer up to 1 kW at efficiencies as high as housing with the electronics necessary to support power and
88%. data transfer between the vehicle, dock and cable is mounted
While we retained FAU’s original design for the prototype below the tube, along with the acoustic transponder used in
demonstration described here, in the future we plan to homing. A 2 M-pixel Ethernet camera is positioned below the
reconfigure the system so that the moving parts and actuator cone (not shown in Fig. 4), looking up at about 45 degrees to
are in the AUV. This change will retain our design allow shore-side operators to monitor the approach, and

0-933957-35-1 ©2007 MTS

confirm the vehicle is present and fully aligned inside the the cross-track error, and also prevents the vehicle from hitting
dock. An important feature of this design is the compliance the dock with too much force. 5) Alignment. The tapering
that is built into the cone that helps absorb energy as the AUV cone and tube align the AUV in all axes except roll and surge.
quickly decelerates as it enters the opening of the tube. In the A V-groove entrance and axial slot cut in the top of the tube
event of a direct hit where the AUV bypasses the compliant force the vehicle’s lifting eye that protrudes from the top of
cone and enters the tube at full speed, a series of vent holes in the AUV, into alignment. A hard urethane bumper at the end
the tube restrict water exiting from the tube, limiting the of the axial slot arrests the AUV’s surge axis. 6) Latch. The
AUV’s deceleration rate before hitting the end-stop. vehicle monitors an inductive position sensor and the status of
the data transfer Ethernet connection to determine when to
A. Gimbaled Base
command the insertion of the locking pin. Once in position,
The seafloor around MARS, the planned deployment site,
the locking latch/charge pin is raised and the vehicle is locked
slopes toward the abyss at about 8 degrees. We plan to deploy
in place. In the open ocean tests, shore operators raised the pin
the dock at this site by lowering it on a cable, touching down
remotely while watching the vehicle enter the dock on the
and firing an acoustic release to separate the dock from the
camera. If the AUV is unable to fully align, it backs out and
lifting wire. As we’d have little control over the orientation,
re-enters the approach pattern to attempt docking again. The
we were concerned the docking process could be complicated
vehicle undocks by lowering the latch/charge pin and
by the tilted tube. Therefore, we implemented a two axis
reversing the thruster. The vehicle is unstable in reverse, and
gimbal between the dock tube and tripod base to allow the
therefore thrusts with a fixed elevator angle and duration.
tube to self-level, Fig 4. A 275 kg railroad wheel at the end of
After the thruster shuts off, the positively buoyant vehicle
the pendulum provides a strong righting moment, leveling the
floats up above the dock and begins a normal mission with
dock. In addition to self- leveling, this system provided
forward thrust. A more detailed description of the control
additional dampening to help reduce accelerations during the
system used can be found in [13].
docking sequence. Since we did our initial testing in 12
meters of water, surface waves rocked the dock, forcing us to
X. SEA-TRIAL RESULTS
disable the gimbal motion with cable stays to the three legs of
the frame. The AUV and docking station were tested in MBARI’s 1.4
million liter seawater test tank before going to sea (Fig. 1).
IV. HOMING AND DOCKING SEQUENCE
Because the MARS observatory was not operational in time
The general homing and docking sequence consists of the for our initial tests, we accepted an invitation from Tim
following steps: 1) Locate and home to the dock. The vehicle Stanton at the Naval Postgraduate School to use the Monterey
uses its on-board navigation to transit within USBL range of Inner Shelf Observatory (MISO). MISO provides power and
the beacon, which is about 2 km. The vehicle then homes to Ethernet connectivity to a node in 12 meters of water about
the dock using pure pursuit guidance, where the heading 600m off the beach near Monterey California. The dock was
control system simply keeps the vehicle pointed at the beacon. installed in early June 2006 and was recovered five months
This guidance strategy does not compensate for ocean current later (Fig. 2). Bio-fouling and wave motion were two
and thus the vehicle can be blown downwind while challenges that we hadn’t anticipated for the planned 900
approaching. Pursuit guidance has the advantage of keeping meter deep deployment on MARS. Based on historical records
the USBL pointed right at the beacon for maximum signal from an upward looking current profiler (ADCP) installed at
strength. 2) Compute a position fix. When the USBL attains MISO we anticipated wave induced reversing currents of 0.25
good signal strength, the vehicle uses its compass heading and m/s, which were within the 0.5 m/s limit we had in mind
the USBL bearing and range to compute a position fix. Based during the development of the AUV docking control system.
on previous testing this occurs between 1000 and 200 meters To reduce the chance of MISO being damaged by fishing
away from the dock. Since the final homing control does not trawlers the instrument frame and connector ports were
use compass heading or propagated position, a jump in the installed in a 100m wide sand channel between two large kelp
propagated position caused by the fix does not affect the beds. We defined navigation waypoints to guide the AUV in
vehicle control. 3) Fly to the start of the final approach path. from open water, down the channel and into the dock. When
The approach path is along the cone centerline and begins the AUV left the dock, or missed an approach, it would make
about 300 meters out. The approach may require the vehicle to a U-turn and head out the channel to a waypoint in open
turn away from the dock and temporarily lose USBL contact. water.
4) Execute final approach. The vehicle approaches along the During the early trials at MISO the AUV made several
cone centerline using a cross-track controller instead of pure unplanned excursions into the kelp due to erroneous data from
pursuit. This means that the vehicle will track the approach the USBL. A moving-window median filter was developed to
path to the cone, and acquire a drift correction angle (crab filter the USBL. A median filter can be effective for sonar
angle) if the ocean current has a lateral component. This type applications because it rejects multi-path errors, as long as the
of control uses both range and bearing from the USBL at 1 Hz, number of measurements corrupted by multi-path is fewer
and uses the DVL and compass to dead-reckon between than the uncorrupted ones.
USBL updates. The vehicle slows to 1.0 m/s about 200 m Fig. 6 shows the vehicle path, as seen from above, of the
from the dock. This allows time for the control loop to zero second docking mission. The axes are eastings and northings

0-933957-35-1 ©2007 MTS

(UTM) in meters. The magenta points at the start, near (0,0),
are GPS hits. During this mission, the vehicle submerges,
turns and flies north to the first waypoint, then turns again to
the south and passes through two more waypoints. The color-
coded diamonds overlaying the path show the location of the
vehicle at each USBL measurement. The color shows the
distance from the transponder, and corresponds to the color of
the measured location indicated by a +, and the black ’s
show the location of filtered points. The second chart in Fig 6
shows a zoom of the dock location. There is a distinct cluster
of points about 8 meters to the west of the dock, detected by
the USBL when the vehicle was around 200 meters out. The
error is presumed to be due to a repeatable bounce (multi-
path). This cluster contains many valid hits. As the vehicle
closes, however, this cluster is discarded by the moving
window and the valid hits then cluster at the dock, precisely as
desired. The position error of the vehicle when it enters the
dock, as viewed on the video and in Fig. 5, is nearly
undetectable, appearing less than 5 cm.
A shift in priorities at MBARI caused an abrupt end to
testing. On the last day of testing we ran four missions and
were able to demonstrate the complete docking sequence on
the first attempt for each of the four missions. Automatic
detection of the dock by the AUV was intermittent, but shore-
side operators were able to watch the AUV fly into the dock,
confirm alignment and manually raise the latch/charge pin.
Power was transferred to the battery at a rate of 416 watts with
a measured efficiency of 48%. Mission data was transferred
off the AUV at 10 Mb/sec and the AUV was completely
powered down and then powered up, and a new mission was
downloaded. The new mission was started, the pin was
lowered, and the vehicle backed out, cleared the dock and ran
the next mission.
XI. CONCLUSIONS
Figure 6: Image above shows a post mission plot showing AUV position track
A prototype docking station for a 21” (54 cm) AUV has on a docking mission, and a zoom in image below that shows the position
been designed, built and tested at sea. The AUV and dock was estimates of dock based on USBL positioning.
ocean tested while connected to the MISO cabled observatory
in Monterey Bay. All the of the steps required for docking a
21” AUV to a seafloor observatory has been demonstrated,
including: 1) homing and capture of the AUV in the docking
station 2) establishing an Ethernet connection between the
vehicle and the dock using a short range radio frequency
communication system 3) complete interactivity with the
vehicle including recovering mission data to shore,
downloading new missions to the vehicle, and even adding
new code to the vehicle and compiling in situ 4) inductive
transmitting power to the vehicle 5) charging vehicle batteries
from observatory power 6) turning the vehicle off, and then
waking it up from the dock 7) undocking the vehicle from the
dock, and commencing missions. MBARI plans to refine this
prototype system and install and operate an AUV for science
missions from the MARS cabled observatory.
ACKNOWLEDGMENT
This work would not have been possible without important
Figure 5: Screen capture image from dock camera showing AUV entering the contributions from Hans Thomas, Rich Henthorn and Steve
dock

0-933957-35-1 ©2007 MTS

Rock (Stanford University). We would particularly like to
thank Tim Stanton (Naval Postgraduate School) for allowing
us full and unrestricted access and use of the MISO, which
provided the undersea power and data cable, and infrastructure
required to test the dock. We gratefully acknowledge the
support of the Packard Foundation. This work was supported
in part by the National Science Foundation under NSF-OPP
9910290. Any opinions, findings and conclusions or
recommendations expressed in this material are those of the
author(s) and do not necessarily reflect those of the National
Science Foundation.
REFERENCES
[1] R. Stokey, M. Purcell, N. Forrester, T. Austin, R. Goldsborough, B. Allen
, C. Von Alt, A docking System for REMUS, an Autonomous
Underwater Vehicle, MTS/IEEE Conference Proceedings Oceans ’97,
Halifax, NS, Canada, pp. 1132-1136.
[2] J. Lambiotte, R. Coulson, S. Smith, E An, Results from mechanical
docking tests of a Morpheus class AUV with a dock designed for an
OEX class AUV, MTS/IEEE Conference Proceedings Oceans ’02 Biloxi,
MS Oct. 2002 pp. 260 - 265
[3] T. Fukasawa, T. Noguchi, T. Kawasaki, and M. Baino, “Marine bird - a
new experimental auv with underwater docking and recharging system,”
in MTS/IEEE Oceans 2003 Conference Proceedings, San Diego, CA,
September 2003, pp. 2195–2200.
[4] H. Singh, J. G. Bellingham, F. Hover, S. Lerner, B. A. Moran, K. Heydt,
and D. Yoerger, “Docking for an autonomous ocean sampling network,”
Journal of Oceanic Engineering, vol. 26(4), pp. 498–513, October 2001.
[5] S. Cowen, S. Briest, and J. Dombrowski, “Underwater docking of an
autonomous undersea vehicle using optical terminal guidance,” in
MTS/IEEE Oceans 1997 Conference Proceedings, Halifax, Canada,
October 1997, pp. 1143–1147.
[6] R. Coulson, J.C. Lambiotte, G Grenon, T. Pantelakis, J. Curran, E. An,
Development of A Modular Docking Sub-System for 12” Class
Autonomous Underwater Vehicles, MTS/ IEEE Conference
Proceedings Oceans ’04, Kobe Japan Nov. 2004pp. 1745- 1749 Vol.3
[7] P. M. Lee, B. H. Jeon, and S. M. Kim, “Visual servoing for underwater
docking of an autonomous vehicle with one camera,” in MTS/IEEE
Oceans 2003 Conference Proceedings, San Diego, CA, September
2003, pp. 2195–2200.
[8] J. Evans, P. Redmond, C. Plakas, K. Hamilton, and D. Lane,
“Autonomous docking for intervention-style auvs using sonar and
video-based real-time 3d pose estimation,” in MTS/IEEE Oceans 2003
Conference Proceedings, San Diego, CA, September 2003, pp. 2201–
2210.
[9] M. Gertler, “Resistance experiments on a systematic series of streamlined
bodies of revolution – for application to the design of high-speed
submarines,” Navy Department, The David taylor Model Basin, Report
C- 297, Apr 1950.
[10] R. S. McEwen and K. Streitlien, “Modeling and control of a variable-
length auv,” in 12th International Symposium on Unmanned Untethered
Submersible Technology, University of New Hampshire, Durham, NH,
August 2001.
[11] M. D. Feezor, F. Y. Sorrel, P. R. Blankenship, and J. G. Bellingham,
“Autonomous underwater vehicle homing/docking via electromagnetic
guidance,” IEEE Journal of Oceanic Engineering, vol. 26, no. 4, pp.
515– 521, 2001.
[12] J. G. Bellingham, R. S. McEwen, and B. W. Hobson, “The development
of a docking station for a regional cabled observatory,” in 4th
International Workshop on Scientific Use of Submarine Cables, Dublin,
Ireland, February 2006.
[13] R. S. McEwen, B. W. Hobson, J.G. Bellingham, Docking Control System
For A 21” Diameter AUV, The Proceedings of UUST 2007, Durham
NH, August 2007.

0-933957-35-1 ©2007 MTS

View publication stats