Abstract—JAMSTEC has developed the two types of
underwater vehicle since 2005: an ROV to the oceans' deepest
depth called ASSS11k(Advanced Sediment Sampling System to
11,000 meter) and a hybrid underwater vehicle for use in
shallow-water to mid-depth zones named PICASSO (Plankton
Investigatory Collaborating Autonomous
The most important purpose of the ASSS11k is to get a lot of
mud sample of Challenger Deep in the Mariana Trench, because a
number of bacteria have been found there. Scientist wants to
continuously explore the deepest parts of the oceans with a vehicle
equipped with sediment samplers. ASSS11k consists of a sampling
station and a sediment probe. The station contains two types of
bottom samplers. One launches the probe to make a preliminary
survey, launching the sampler to obtain a sample. We carried out
the first sea trial using support vessel of “KAIREI” in January
2007. We tested every functions of the system and achieved
sediment sampling at Sagami bay.
PICASSO (2 x 0.8 x 0.8 m, 200 kg) is designed for biological
and oceanographic observations in depths of up to 1,000 m. This
small, light vehicle can be handled and operated by a team of only
a few people. The easy-to-use vehicle does not need a dedicated
support ship. The vehicle can be used either as untethered
remotely operated vehicle (UROV) or autonomous underwater
In order to develop these vehicles, we used some new
technologies and then developed new original devices: a small
electrical-optical hybrid communication system, an HDTV optical
communication system with Ethernet interface, synthetic
designed pressure vessel- chassis- inner circuit boards, buoyancy
material for deepest depth, a thin cable with high-tensile strength,
a core sampler launcher, crawlers, compact winch motor drivers,
a USBL system, a ballast controller, a friendly-user-interface
program for operator, a high capacity lithium ion battery, a down
sizing optical fiber spooler, and a prototype of underwater
electromagnetic communication system.
Index Terms—Autonomous Underwater Vehicle, Remotely
Manuscript received April 2, 2007.
Hiroshi Yoshida, Taro Aoki, Hiroyuki Osawa, Satoshi Tsukioka, Shojiro
Ishibashi, Yoshitaka Watanabe, Junichiro Tahara, Tsuyoshi Miyazaki,
Tadahiro Hyakudome, Takao Sawa, Kazuaki Itoh are with Marine Technology
Center, Japan Agency for Marine-Earth Science and Technology, Yokosuka,
Japan (corresponding author to provide e-mail: firstname.lastname@example.org).
Akihisa Ishikawa is with Nippon Marine Enterprises, Ltd.
Dhugal Lindsay is with Extremobiosphere Research Center, Japan Agency
for Marine-Earth Science and Technology, Yokosuka, Japan.
AMSTEC is owner of the manned submersible SHINKAI
6500, the 3,000 m-class remotely-operated vehicle
HYPER-DOLPHIN, and KAIKO 7000II  and has
operated them. Scientists use these vehicles in accordance with
their research objectives in depth of up to 7,000 meters. Every
vehicle has high performance and is very useful, but is not
sufficient to investigate whole deepsea.
Recently, a number of bacteria have been found from mud
samples of Challenger Deep in the Mariana Trench . Those
sediment samples were taken with KAIKO which was the only
ROV could reach to the deepest depth, but the vehicle of the
KAIKO was lost -.
In order to investigate the distributions of macro- and
micro-plankton versus environmental parameters, several trials
with ROVs and manned submersible  have been carried out.
In these ways, one is only able to gain the information of a point
nature and not be able to determined large scale distributional
patterns with limited ship-time.
As mentioned above, some microbiologists want an ROV
reached to the deepest depth to get mud samples. Some
scientists need small underwater vehicles can be operated
without a dedicated support ship for research of midwater zone.
It is difficult and may cost so much to develop an underwater
vehicle which provides both requirements. We have thus
started design, development and construction of two types of
vehicles in 2005: an 11,000 m-class ROV system for sediment
sampling and a small multiple-platform autonomous survey
system which is deployable from small to medium sized boats
and ships of opportunity. The development for each system so
far costs about 1 million dollars, respectively.
The ROV system equipped with a sediment sampling tool
will dive up to 11,000 meter so that the system must enable
scientist to get sediment samples in the deepest ocean floor. We
named this ROV system as ASSS11000 (Advanced Sediment
Sampling System 11000). The ASSS11000 is mounted on the
support vessel KAIREI. The KAIREI is already equipped with
the on-board system for the KAIKO7000. We thus divert some
part of the on-board system of the KAIKO7000 to the ROV to
cut its production cost.
The other vehicle system is based the small hybrid
underwater vehicle (UROV/AUV), MROV  developed by
JAMSTEC and are equipped with a high resolution camera
Newly-developed Devices for The Two Types of
Hiroshi Yoshida, Taro Aoki, Hiroyuki Osawa, Satoshi Tsukioka, Shojiro Ishibashi, Yoshitaka
Watanabe, Junichiro Tahara, Tsuyoshi Miyazaki, Tadahiro Hyakudome, Takao Sawa, Kazuaki Itoh,
Akihisa Ishikawa, and Dhugal Lindsay, Member, IEEE
1-4244-0635-8/07/$20.00 ©2007 IEEE
system, a VPR and environmental sensors. The survey system
consist of the multiple small underwater vehicles with a 1,000m
depth rating, working in concert, will overcome all of these
previous shortcomings. The use of JAMSTEC’s 1.5m2
IONESS net and an ROV or HOV with specimen sampling
capabilities will enable calibration and ground-truthing of the
data collected by the vehicle system.
We introduce outline of the vehicles and then describe the
newly-developed devices installed
Communication systems especially are focused. Finally, some
results of the sea trial will be given.
in the vehicles.
II. OUTLINE OF THE TWO VEHICLES
A. A 11,000 m class ROV for Sediment Sampling
The ASSS11k and its 11,000m-cable store winch are
mounted on the dedicated ship “KAIREI”. The ASSS11k
consists of an on-board equipment which is installed in the ship,
a sampling station, a sediment probe, and two samplers. Fig. 1
shows a recovery scene of the station housed the probe. The
on-board equipment is connected with the station via the
primary cable. The secondary cable newly developed connects
between the station and the probe. The sampling station houses
the probe and one of the samplers in the bottom cage. The
station is mounted a docking-undocking system and a
secondary cable drum for the probe, and a sampler release gear
and a rope-hoisting winch for the sampler. The station
furthermore serves as repeater between the on-board equipment
and the probe. The probe cruises under around the station freely
within the reach of the 160 m cable to survey sea-bottom
surface with a TV camera. The probe is able to take small
amount of a sample of sediment with a mini manipulator. Two
types of sediment samplers - a gravity core sampler and a grab
bottom sampler are prepared. Scientist can choose either
sampler in accordance with the intended use. The system
specification is shown in TABLE I.
B. A small vehicle for midwater plankton survey
The Vehicle of the system must be small because the vehicle
does not need a dedicated support ship to be unconstrained by
limited ship-time. Multiple vehicle configuration and
simultaneous deployment of the vehicle are needed because
small vehicle is not able to have many instruments for
observations and does not cruise long distance. Multiple
configuration can be covered these disadvantages of the small
vehicle. From above requirements, we have designed the
vehicle with the following things: light weight (under 200 kg),
small size (approximately 2 m), long duration (over 5 hrs),
1000 m depth ratings, having a HDTV camera or Visual
Plankton Recorder, semi-autonomously detecting and tracking
a plankton. We plan to make the system consists of two or more
Since 2006, we have developed a small vehicle named
“PICASSO”, Plankton Investigatory
Autonomous Survey System Operon. Fig.2 shows a snap shot
of PICASSO in a sea trial. PICASSO is small and light (2.4 m
long, 200 kg in weight) and of body color is mostly red because
plankton hardly recognizes its wave length.
The vehicle system consists of an on-board equipment and a
vehicle which are connected via a thin optical fiber cable. One
remotely controls the vehicle from the equipment. PICASSO is
composed of major parts: an FRP fairing cover, a body frame,
buoyancy materials, controllers, communication systems, three
100 W thrusters, one tilt actuator, lights, devices for navigation
and observation, oil-filled lithium ion battery, and an optical
fiber spooler. The vehicle has one vertical tail fin and two fins
for stability. Table 1 shows the PICASSO specifications.
One of the purposes for development of the system is to track
FIG. 1. AN OVERVIEW OF THE ASSS11K. THE ROV FINISHED A MUD
SAMPLING, BEING RECOVERED.
SPECIFICATIONS OF THE ASSS11K
ITEMS THE STATION THE PROBE
11,000 m 11,000 m
Size 2.1 x 2.7 x 3.0 m 1.3 x 1.1 x 1.2 m
Weight 2,000 kg 300 kg
gravity core sampler /
grab bottom sampler
TV camera, CTD*
φ 45 mm x 12,000 m
2 aft, 2 vertical,
φ 20 mm x 160 m
Fig. 2. An overview of PICASSO just recovered.
plankton. A typical velocity of planktons is slower, a few
hundreds meter per hour , and its direction of swimming is
random. PICASSO, therefore; has two lateral thrusters with
180 degree tilter and a vertical thruster for highly maneuverable
cruising and its maximum cruising speed is set to 2 knots. For
smaller to larger animal observation PICASSO is able to select
a main imaging tool from among three choices that are a HDTV
camera, 12 bit high resolution camera, and the microscope
SPECIFICATIONS OF PICASSO
III. DEVICES FOR COMMUNICATION
A. An optical-electrical communication system
For two different types of vehicle, two kinds of
communication system were developed. One is an
optical-electrical communication system for the ASSS11k and
the other is a high speed one for PICASSO.
In the ASSS11k system, a 3-point-communication (the ship –
the station – the probe) is needed. The block diagram of the
optical communication system model-numbered “JT3” is
depicted in Fig. 3. For communication between the ship and the
station, an optical communication is utilized but the probe is
communicated with the station via a metallic coaxial cable
because of cost down. It is technically very demanding and
highly costly to produce an optical slip ring fixed to a cable
drum used in depth of 11,000 meters. Its optical communication
bit rate is the same as the SONET (STM-4) standard but the
protocol is an original one. Every input signal is sampled, time
shared, Manchester encoded, and then transmitted at a bit rate
of 622 Mbps. The radio frequency digital communication
device, “JT3-RC” for the station-probe communication is a full
duplex transceiver with 8 RS-232C ports and its carrier
frequency is about 10 MHz. In JT3-RC circuit board, its
synchronization is achieved by a sequential synchronization
using Manchester encoding with 16 bits preamble. The
time-division multiplex data rate is 12.96 Mbps. Maximum
transmission range is designed to be 200 meters by using
2.5-2V standard coaxial cable. A pre-emphasis circuit reduces
deform of the transmission wave caused by loss of the cable. In
a land test of the system, it was found that the motor for the
thruster generates electric noise and this noise affects the
JT3-RC to often occur communication error. In this year, we fix
this problem by using Injection Locked Oscillator which can
reject the environmental noise and speed up regeneration of
B. A 3Gbps optical communication system with HD-SDI
In order to transmit an HDTV signal, a communication device
was developed. The wideband device is interfaced an operator
to the vehicle, having five interfaces: one HD-SDI data port is
for an HDTV camera, three NTSC ports, an 100 Mbps Ether net
port, four RS-232C ports and two RS-485 ports. The data from /
to these ports are serialized / deserialized by TLK3101 giga bit
transceiver. A 2488 Mbps bit rate optical transceiver module
produced by Sumitomo Electric Industries, Ltd. then converts
electrical signal to / from optical signal. The custom made
optical communication device installed in the vehicle is small
(three printed circuit boards of 120 x 80 mm).
IV. THE OTHER DEVICES
A. Secondary Cable
We have developed a new
cable using para-aramid fiber
with the tensile strength of
350kg/mm2 since 2005. This rod
type aramid fiber does not make
stress concentration. The cable
(φ20 mm x 160 m) consists of
this aramid fiber, two coaxial cables, four single wire cables for
power lines, cable sheath, and resin. The cable is covered
polypropylene clothing. Specific gravity of the cable is around
1.3 and rupture strength is about 70 kN.
2.1 m x 0.8 m x 0.8 m
UROV / AUV
2 horizontal 100 Watt thrusters
with tilt system,
1 vertical 100 Watt thruster
2 G bps optical communication
Radio LAN, ARGOS
transmitter, acoustic and
MEMS gyro, Doppler velocity
log, depth meter, SSBL,
4 x NTSC cameras, 3 x 35 Watt
HID lamps, Visual plankton
recorder*, High definition TV
camera*, Digital still camera*,
400 watt HID lamp*.
* install any one
of the devices.
* under development.
Fig.3 (a). Blockdiagram of the optical communication part of the JT3.
Fig. 3 (b). Blockdaiagram of the JT3-RC. The synchronizer regenerates
Fig. 4. A prototype of the
B. Buoyancy materials
New buoyancy material is developed. Prototype one is
applied to the ASSS11k. The specifications of the prototype are
crush pressure of 56 MPa and specific gravity of 0.63.
C. Ultra Short Base Line System
The system consists of two major parts: a USBL transceiver
installed on the station and a transponder fixed on the probe.
Table III shows the specifications of the USBL system. The
accuracy of the position is relatively low because the probe
position is directly obtained with the station TV camera in our
plan. In this system, M-sequence signal is used as modulation
signal. We have developed an original processing board on
which a DSP (Black Fin, Analog devices) and an FPGA
(Cyclone, Altera) are mounted. Whole source codes are
developing in JAMSTEC. Four hydrophones have been also
D. Acoustic transceiver
An advanced acoustic communication method utilized
time-reversal waves has been developed. In most acoustic
communications ship-vehicle configuration is vertically
because there many multi-path signals in horizontal
configuration. It is the best way to use time-reversal technique
for communication under multi-path fading. Shimura had made
a simulation for communication between a ship and a vehicle in
shallow water zone using high frequency.  He reported that
the method of time-reversal process with an adaptive filter
provides good communication result. When the vehicle,
however; moves, the advantage of the method is depressed. We
will try to modify the method and choose the best parameters,
aiming better ship-vehicle communication up to 500 m in
E. Communication by magnetic field.
We have been developing a new communication tool which is
applied electromagnetic wave. This method is used for mutual
communication between vehicles up to 50 m. We made a
prototype transmitter, a receiver, and antennas. An NTSC
camera for underwater is connected to the transmitter. The
transmitter encodes and modulates the image data and then
supplies power of 17 Watts to a multi-turn coil antenna. A high
sensitivity search coil antenna receives the data modulated. The
receiver demodulates, decodes, and outputs the image with
QVGA format. In the tank test, QVGA image was transmitted
to the receiver set 30 m apart from the transmitter. We must
carry out the test in the sea.
V. MIDWATER ANIMAL TRACKER
PICASSO will semi-automatically track animals in
midwater. In order to detect and track an animal the vehicle has
the function of animal image recognition and then
automatically moves not to lose the animal recognized. We
have developed a prototype system of the animal tracker using
MROV. To simplify the prototype system, the pan-tilt system
of the camera is only controlled by a tracking program. Color
deference in HLS (Hue, Saturation, Luminance) color space is
basically used for detection. Identification of a target is initially
done to click the target on the display. Values of RGB in the 9 x
9 pixels around the pixel clicked are converted to of HLS. A
center of gravity of the pixels which have near-HLS value
obtained is then calculated. When the distance between the
center of gravity and the center of the image obtained by the
camera exceeds a limit preset of the pan-tilt, the program
controls the camera to center the animal. The program also has
a function of the displacement prediction of an animal.
A detection and tracking test was carried out in the large fish
tank (6.5 m in depth, 144 m2 area of base) in Enoshima
Aquarium. A scene of test is shown in Fig.5. The prototype
system have detected and tracked a small fish (red circle in the
figure) for 30 seconds in this test. But in most case the duration
of capturing is a few seconds because there are many fish in the
tank and background-target contrast is low compared to in
midwater. For more accurately detection, we will collaborate
on an image recognition method with MBARI . This
method simulates human vision function and has high target
recognition probability. We furthermore investigate the
program to track with thruster control.
TARGET SPECIFICATIONS OF THE USBL.
Accuracy <5% within 200 m range
Range 2,000 m
Depth ratings 11,000 m
Frequency 20 kHz
Data M-sequence signal
Sensors Sound velocity meter
Transducer 4 array
TX sound pressure 180 dB re uPa at 1m
RX sensitivity -210 dB re 1V/uPa at 1m
Fig. 5. A test of tracking using the MROV in the Enoshima Aquarium.
VI. SEA TRIAL RESULTS
A. The sea trial of the ASSS11k in January 2007
We have carried out a sea trial in 5 – 9 January in Sagami
Bay using the support vessel, KAIREI. We made three dives
(see TABLE IV) for 5 days because of bad weather. The test
area was limited to keep away from the strong wind.
First dive was made at Yokosuka 4th district in Tokyo bay to
check system function on January 5. The test results were good
and we then headed to Sagami Bay. It was determined in
second dive whether the station thruster can constrain its self
rotational motion caused by the primary cable twisting or not.
The thrusters behave well up to 200 meters dive. On January 8
we tried sampling with the gravity core sampler at the location,
which bottom sediment is softish, with depth of 480 m. The
station was controlled keeping its heading, coming as close as
about 80 meters to the bottom. Keeping its altitude, the sampler
was dropped. The sampler was recovered after about ten minute
later. We got sediment sample about 200 mm long as shown in
Fig.6. The sample obtained is under analysis by microorganism
B. The sea trial of PICASSO February-March 2007
We have carried out a sea trial in 24 February – 4 March in
Sagami Bay and Suruga Bay using the support vessel,
NATSUSHIMA. We made seven dives. In two dives of them,
the VPR was installed on PICASSO (Fig.7). The vehicle dived
to depth of up to 601 meters. All function tests were performed
well. The image obtained by the HDTV was wonderful and
some animals are observed. The film will be presented in the
conference. A few mill meters planktons were observed by the
VPR (Fig. 8). This first sea trial of PICASSO was successfully
We introduce the two types of the vehicles; the ASSS11k for
sediment sampling at the deepest depth and PICASSO for
survey plankton in midwater zone. The optical communication
systems developed for the both vehicles are described. A high
tensile cable, a buoyancy material, an acoustic positioning
system, wireless communication systems, and plankton
tracking system are also described. The sea trial results of the
vehicles installed above devices are reported. We improve
these devices through testing, completing the both vehicles.
We would like to thank for Enoshima Aquarium to
collaborate the detection and tracking test and for KOWA
Corporation to help development and test of the vehicle. The
author also thank for GRAVITON Inc. to manufacture both
 Murashima, T, Nakajoh, H, Yoshida, H, Yamauchi, N, and Sezoko, H,
“7000m class ROV KAIKO7000,” Proc. OCEANS ’04 MTS/IEEE, pp.
812 - 817.
 Nakajoh, H, Murashima, T, and Yoshida, H, “7000m operable deep-sea
ROV system, ‘KAIKO7000’ ,” Proc OMAE, 2005.
 Takai, K, Inoue, A, Horikoshi, K, “Thermaerobactor marianensis gen.
nov.,sp. nov., an aerobic extremely thermophilic marine bacterium from
the 11000 m deep Mariana Trench,” Int J Syst Bacteriol 49 (2) pp.
 Momma, H, Watanabe, M, Hashimoto, K, and Tashiro, S, “Loss of the
Full Ocean depth ROV Kaiko – Part 1: ROV Kaiko – A Review, ” Proc.
of 14th ISOPE-2004, pp.191-193.
 Tashiro, S, Watanabe, M, and Momma, H, “Loss of the Full Ocean depth
ROV Kaiko – Part 2: Search for the ROV Kaiko Vehicle, ”Proc. of 14th
 Watanabe, M, Tashiro, S, and Momma, H, “Loss of the Full Ocean depth
ROV Kaiko – Part 3: The Cause of Secondary Cable Fracture,” Proc. of
14th ISOPE-2004, pp.199-202.
THE CONDITION FOR DIVES
30 m Tokyo bay
200 m Sagami bay
400 m Sagami bay
No vertical lines in table. Statements that serve as captions for the
entire table do not need footnote letters.
Fig. 6. A mud sample obtained by the ASSS11k at Sagami
Fig. 7. An overview of PICASSO having the Visual Plankton
Recorder just recovered.
Fig. 8. A jelly fish and a GNATHOPHAUSIA obtained by the Visual
Plankton Recorder installed in PICASSO.
070131-018 Download full-text
 Wiebe, P. H. and Benfield, M. C. (2003). “From the Hensen net toward
four-dimensional biological oceanography,” Progress in Oceanography
56, pp. 7-136.
 Yoshida, H, Tsukioka, S, Hyakudome, T, Ishibashi, S, and Kitamura, M
(2005). “A new study platform for plankton – Autonomous underwater
vehicle –based observation,” Bulletin of the plankton society of Japan,
Vol.52, No. 2, [in Japanese].
 Matsumoto, G. I. (1991). “Swimming movements of ctenophores, and
the mechanicsof propulsion by ctene rows,” Hydrobiologis 216/217, pp.
 Shimura, T, Yoshida, H, and Lindsay, D (2006). “Basic study on
time-reversal communication in shallo water with high frequency,” Proc.
of Techno-Ocean 2006 / 19th JASNAOE OES. Paper No. 171.
 Walther, D., Edgington, D.R., Koch, C., (2004). “Detection and tracking
of objects in underwater video,” Proceedings of the 2004 IEEE Computer
Society Conference on Computer Vision and Pattern Recognition, Vol. 1,
pp. I-544- I-549 Vol.1.