A Customisable Underwater Robot PDF
A Customisable Underwater Robot PDF
Abstract—We present a model of a configurable underwater operated vehicle 1 ) and BlueRov2 (an opensource ROV kit
arXiv:1707.06564v1 [cs.RO] 20 Jul 2017
drone, whose parts are optimised for 3D printing processes. We produced by BlueRobotics 2 ).
show how - through the use of printable adapters - several Not surprisingly, research institutes and do-it-yourself com-
thrusters and ballast configurations can be implemented, allowing
different maneuvering possibilities. munities exhibited a growing interest in the adoption of
After introducing the model and illustrating a set of possible opensource underwater drones for many applications [3]: from
configurations, we present a functional prototype based on open underwater exploration to marine science, archeology and
source hardware and software solutions. The prototype has been marine geoscience; from diving to fishing; from boating to
successfully tested in several dives in rivers and lakes around harbour and boat maintenance and support; from garbage
Berlin.
collection to operation in hazardous environment.
The reliability of the printed models has been tested only
in relatively shallow waters. However, we strongly believe that
their availability as freely downloadable models will motivate
the general public to build and to test underwater drones,
thus speeding up the development of innovative solutions and
applications.
The models and their documentation will be available for
download at the following link:
https://adapt.informatik.hu-berlin.de/schillaci/underwater.
html.
I. I NTRODUCTION
Underwater robots are tipically classified into two cate-
gories: Autonomous Underwater Vehicles (AUVs) and Re-
motely Operated Vehicles (ROVs). AUVs are unmanned, self-
propelled vehicles that can operate autonomously of a host
vessel, which makes them well suited to exploration of ex-
Fig. 1. A screenshot of a 3-thrusters prototype.
treme environments [1]. A continuous development of new
vehicles and sensors, and advances in artificial intelligence
Similarly to flying drones, low cost underwater vehicles are
are increasing the range of AUVs applications [2]. However,
usually equipped with cameras and basic sensing capabili-
their high cost and the know-how required to develop them
ties (e.g. inertial measurement unit, depth sensor, etc.), and
makes their adoption prohibitive to the general public.
are teleoperated from the surface. Underwater propulsion is
Remotely operated underwater vehicles are mobile devices
achieved through propellers, whose number and configuration
commonly used in deep water industries. ROVs are usually
(e.g. with fixed or changing orientation) define the degrees
connected to a surface vessel through a tether and controlled
of freedom of the movements that the drone can perform
by an operator, which makes them more reliable and less prone
underwater. A conventional procedure in ROVs/AUVs design
to failures than AUVs. Similarly to AUVs, professional ROVs
is to make vehicles positively buyoant to ensure they will
are tipically very expensive.
return to the surface, should any power failure occurs. More-
However, thanks to the rapid growth of opensource software over, low-cost ROVs adopt fixed ballast - achieved through
and hardware communities, and thanks to their experience lead, buoyancy chambers and syntactic foam - to have the
in the development of flying drones, underwater teleoperated desired gravity. Under a fixed ballast configuration, ascend
drone technology has recently become accessible to a wider and descend movements are achieved using vertical thrusters.
public. In fact, low-cost open-source solutions for remotely
operated vehicles (ROVs) can be found in the market since 1 https://www.openrov.com/
Fig. 5. The split adapter can be used to split a mounting point into two. In
the figure, a combination of split and 90◦ adapters is used to mount a vertical
and a 45◦ horizontal thrusters onto the same mounting point in the enclosure
clamp.
Fig. 3. The three adapters that allow setting different orientations of the
thursters and ballast.
Fig. 4. The 90◦ adapter can be used to mount a vertical thurster onto the
enclosure clamp.
The adapters, thrusters and ballasts can be mounted together Fig. 9. Underwater drone with a 3-thrusters frame configuration.
using standard M3 nuts and M3-30mm bolts.
The parts described above allow multiple frame config-
urations, each providing different maneuvering possibilites
and degrees of freedom for the drone movements. Figure 9
shows perhaps the simplest setup with a 3 thrusters frame
configuration. Here, two horizontal thrusters allow for for-
ward/backward movements and yaw rotations (left and right
turns). A vertical thruster mounted onto the tail mounting point
of the drone provides pitch rotation possibilities (nose up and
nose down movements). As the only vertical thruster is placed
far away from the center of mass of the robot, ascent and
descend movements are not possible with such a configuration.
Although ballast weights can be mounted to adjust the po-
sition of the center of mass, if ascend and descend movements
are required, additional vertical thrusters can be mounted
onto the posterior enclosure clamp (see Figure 10) or onto
Fig. 10. The drone with a 4-thrusters frame configuration. Vertical propellers
a split adpater at the anterior enclosure clamp together with are closer to the center of mass, allowing ascend and descent movements.
a 45◦ thruster (as in the vectored frame configuration depicted
in Figure 11). This configuration may still require adjustments
of the center of mass position through ballast, which can be with a Quad Core 1.2GHz Broadcom BCM2837 64bit CPU
plugged at the bottom of the posterior clamp or at the tail and 1GB RAM - as a companion computer to an opensource
mounting point. flight controller, namely PX4 Pixhawk, sold by mRobotics.
The vectored frame configuration depicted in Figure 11 Pixhawk is an open-hardware project aiming at providing high-
allows for full control on the robot: forward, backward and end autopilot hardware to the public at low costs 4 . The board
lateral movements; ascend and descend movements; yaw and is equipped with several sensors, including a gyroscope, an
roll rotations. accelerometer, a magnetometer and a barometer.
We installed three Turnigy DST-700 brushless motors, each
III. A FUNCTIONAL PROTOTYPE
initially connected to the Pixhawk flight controller through a
In this section, we present an implementation of a functional N-channel MOSFET Afro 12A Electronic Speed Controller,
prototype using a three thrusters frame configuration, as de- which provides PWM control of the motor speed. As we
picted in Figure 9. Section III-A describes the electronics of experienced a damage in a ESCs connected to a propeller that
the drone. In section III-B we describe the software adopted for was put too much under torque stress, we opted for mounting
controlling the robot. We carried out several dives in different Afro 20A ESCs instead of 12A ones.
lakes and rivers around Berlin. Figure 12 shows the drone A Turnigy 5000mAh 4S 25C Lipo Battery Pack is adopted
emerging from a dive in the Dahme river in Köpenick, Berlin. to power the system. Power is distributed to the ESCs through
A. Hardware description a Matek power distribution board, to the Pixhawk through a
mRobotics Power Module and to the Rasperry Pi through a
We adopted low-cost and opensource, when available, so- 5V 3A Switching Power UBEC.
lutions for implementing the electronics of the drone. In
particular, we set up a Raspberry Pi 3 model B - equipped 4 https://pixhawk.org/
from the BlueRobotics online store. Motor power cables and
tether are inserted into the watertight enclosure though cable
penetrators and sealed through J-B Marine Weld marine epoxy.
We acquired also a BlueRobotics waterproof switch that allows
turning on/off the drone without the need of opening the
watertight enclosure. The switch is connected upstream to the
electronics. A pressure valve has been installed for vacuum
seal tests and for simplifying opening/closing the watertight
enclosure.
We acquired also a BlueRobotics Bar30 high resolution
waterproof pressure and temperature sensor. The sensor comes
in a watertight penetrator, thus it can be safely used and
connected into the Pixhawk board.
B. Software description
We rely entirely on open-source software solutions for
controlling the drone. In particular, we adopted Raspian as
operating system for the Raspberry Pi 5 and we installed
ArduSub 6 onto the PX4 Pixhawk. ArduSub is a fully-featured
Fig. 11. Vectored 6-thrusters frame configuration. This configuration allows open-source solution for remotely operated underwater ve-
for better control on the robot: forward, backward and lateral movements; hicles, initially developed by BlueRobotics as derived from
ascend and descend movements; yaw and roll rotations. the ArduCopter code, now is part of the ArduPilot project.
ArduSub has several pre-installed functionalities, including
feedback stability control, depth and heading hold, and au-
tonomous navigation using pre-coded trajectories.
MAVlink is adopted as a communication protocol.
QGroundControl is installed into the surface computer, tablet
or smartphone for providing dive control and mission planning
to the drone. QGroundControl, initially developed as the
control station software for the Pixhawk Project, is an open-
source application used at several research centers and by a
growing communitiy 7 for control of flying, terrestrial and
underwater drones.
Finally, the open source SimonK firmware 8 has been
compiled and deployed onto the Afro ESCs using an Afro USB
programming tool, in order to allow forward/reverse rotations
of the propellers and to match the parameters configured in
ArduSub.
Fig. 12. A picture of the drone during a dive in the Dahme river in Köpenick,
Berlin.
IV. C ONCLUSIONS
In this paper, we presented open source models for building
A Raspberry Pi V2.1, 8 MP 1080P camera module is low-cost underwater drones and for customising their thrusters
installed on the Rapsberry Pi to stream HD videos to the and ballast configurations. The models are optimised for rapid
surface computer. A Lisiparoi low-light led module is installed prototyping through 3D printers.
onboard. The drone is connected to a surface computer through After introducing the different parts and adapters that can
a 25 meters neutrally buyoant RJ45 tether which provides data allow for multiple frame configurations and, thus, different
communication. The user can teleoperate the drone using a grades of maneuverability, we presented a functional pro-
game controller. We tested a XBox360 controller and a ASUS totype. In particular, we built a three-propellers underwater
TV500BG Bluetooth gamepad. Cheap cardboard VR goggles drone based on open source hardware and software solutions.
has been used to stream the video recorded from the Raspberry The prototype has been successfully tested several times in
Pi camera onto a smartphone to ease the teleoperation of the dives near Berlin.
drone.
5 https://www.raspberrypi.org/downloads/raspbian/
Electronics need to be put into a watertight container. We 6 https://www.ardusub.com/
acquired a 4” watertight enclosure, a compatible aluminium 7 http://qgroundcontrol.org/about
end cap and a glass dome cap, and o-ring sealed flanges 8 https://github.com/sim-/tgy
Although the reliability of the 3D printed parts have been
tested only in relatively shallow waters (<10 meters), we
strongly believe that making them freely available online will
motivate the general public to build and to experiment with un-
derwater drones, and consequently speed up the development
of innovative solutions and applications.
Nonetheless, we are planning to carry out further tests in
deeper waters and under different frame configurations. More-
over, we are currently exploring possibilities for autonomous
learning of robot control strategies.
R EFERENCES
[1] R. B. Wynn, V. A. Huvenne, T. P. L. Bas, B. J. Murton, D. P. Connelly,
B. J. Bett, H. A. Ruhl, K. J. Morris, J. Peakall, D. R. Parsons, E. J.
Sumner, S. E. Darby, R. M. Dorrell, and J. E. Hunt, “Autonomous
underwater vehicles (auvs): Their past, present and future contributions
to the advancement of marine geoscience,” Marine Geology, vol. 352,
pp. 451 – 468, 2014, 50th Anniversary Special Issue. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0025322714000747
[2] J. J. Leonard and A. Bahr, Autonomous Underwater Vehicle Navigation.
Cham: Springer International Publishing, 2016, pp. 341–358.
[3] B. Siciliano and O. Khatib, Springer handbook of robotics. Springer,
2016.
[4] H. Hanff, P. Kloss, B. Wehbe, P. Kampmann, S. Kroffke, A. Sander,
M. von Einem, J. F. Bode, F. Kirchner, and M. B. Firvida, “Auvx-a novel
miniaturized autonomous underwater vehicle,” Oceans’17 MTS/IEEE,
2017.