Innovation Report UMV
Innovation Report UMV
net/publication/361999785
CITATIONS READS
0 167
1 author:
Robo-Tech Team
Alexandria University
1 PUBLICATION 0 CITATIONS
SEE PROFILE
All content following this page was uploaded by Robo-Tech Team on 14 July 2022.
Fig1.LEVIATHAN
Table of Contents
Abstract ______________________________________________________________________________ i
Introduction ___________________________________________________________________________ 1
Mechanical System______________________________________________________________________ 2
Electric Cables _______________________________________________________________________________ 2
Electronic Housing ___________________________________________________________________________ 2
Software ______________________________________________________________________________ 2
Bottom-side hosting __________________________________________________________________________ 3
Hardware _____________________________________________________________________________ 3
DC-DC converters ____________________________________________________________________________ 3
Inside the electronics housing __________________________________________________________________ 4
Anti-leakage system __________________________________________________________________________ 4
Image processing _______________________________________________________________________ 4
Machine learning ____________________________________________________________________________ 4
Measuring Objects ___________________________________________________________________________ 4
Introduction
Remotely operated vehicles, or "ROVs," are unmanned, maneuverable underwater vehicles that may be
used to explore the depths of the ocean while being controlled from the surface of the sea. ROVs were
originally created for industrial uses such as internal and exterior inspections of undersea pipelines and
structural testing of offshore platforms, but they are now employed for a wide range of applications, many
of which are scientific in nature. They have proven to be incredibly useful in ocean exploration and are also
utilized for instructional activities at aquaria and to connect to scientific excursions that are taking place in
real time over the Internet.
These underwater robots are often operated by a person on a surface vessel through a joystick. The ROV is
connected to the ship by a tether, which sends electrical signals back and forth between the operator and
the vehicle. Most ROVs have cameras, and lighting allowing them to transmit photos and video back to the
ship. To gather samples, vehicles may be outfitted with additional equipment such as a manipulator and
devices that monitor factors such as pressure and temperature.
In this report, we demonstrate the newest innovative enhancements to Levi, Robo-Tech’s newest ROV, a
light work class ROV equipped with five cameras, two flashlights, one pneumatic gripper, one DC gripper,
two lasers, and several sensors like pressure and temperature. It can fit within a 60cm circle and is
intended to move underwater at the speediest pace with the least amount of drag. Six T-200 Blue Robotics
Thrusters are employed, with two for vertical movement and four for horizontal movement. Now, let's dive
right into our cutting-edge new features for Levi.
Mechanical System
Earlier this year, Orca, our previous ROV, had its sealing compromised by a human diver during a mission,
this pushed us to overhaul our sealing system to be much more robust. So as a result:
Electric Cables
Instead of using pneumatic fittings covered by Teflon as
shown in fig (2) to seal the electric cables going in and
out of the control housing as we did for several years,
this year we decided to use cable glands fig (3), which
also act as a strain relief, giving an extra benefit while
preforming the sealing purpose to the fullest.
Electronic Housing
The sealing used starts with an acrylic housing which is set on
the Polyamide base. Then, in an Aluminum ring, three O-rings
are used between the Acrylic and Polyamide to prevent
water leakage, and the Aluminum cup is fastened by a nut
that pushes on the O-ring fig (4). This year a rubber flange fig
(5) only was used, which performed the same task of sealing
as before without using the previous complicated method.
Software
Remotely Operated Underwater Vehicles (ROVs) suffer the issue of instability most often due to causes
that include but are not limited to drag force on the vehicle and from the tether, misalignment of center of
pressure and center of thrust, inequality of weight force, bouncy force, and moment of Inertia of the ROV.
Most of these cannot be corrected while manufacturing. As a result, an active stability system was
proposed for the counter which uses PID controllers (proportional, integral, and derivative) in three-axis
pitch, yaw, and heave. The system uses our IMU sensor absolute orientation readings, as well as the water
pressure sensor reading and thruster speeds, which are set by the loop to follow a setpoint, which is
updated by pilot controls or the various computer vision algorithms.
Bottom-side hosting
An effective processing unit is required to fully utilize the open-source Robot Operating System (ROS). The
Single Board Computer (SBC) Raspberry Pi 4 functions as the MCU's heart, providing a robust Linux
platform to host the Robot Operating System (ROS), providing the tools, libraries, and capabilities needed
to achieve optimized on-robot autonomy, back-end management, and Interfacing. More specifically, it is
an open-source middleware and a collection of associated software frameworks that allow for modularity
of adjustment, isolation between our programs, and abstraction in design. Our ROS Network fig (6) can be
interfaced with in multiple ways. Either through our user interface, hosted on the onboard Raspberry pi
and connected to the network through ROS-Bridge. Which is responsible for all manual control messages,
or with the topside run Computer Vision Suite, which employs the same topics to carry out autonomous
mission tasks.
Hardware
DC-DC converters
Two sturdy Murata Converters that were chosen for their many significant qualities that strengthened our
system are installed on our power distribution board. Temperature sensing, under-voltage shutdown, and
over current and over voltage protection are some of their amazing features in addition to their smaller
size, which helped decrease the housing size by 10%, compared to their previous counterparts. These
incredible converters also have load-sharing capabilities among themselves, distributing the load equally. A
small heatsink is designed to attach to our all-aluminum housing, an exceptional solution for all our
previous hardware temperature-related issues making this one of our best control tube designs. Using
water as a coolant, our housing is converted into a giant heatsink and this modification made our housing
go beyond sealing solemnly.
Inside the electronics housing
Consequently, cables and electronics speed Raspberry
controllers (ESCs) caused the interior of our PI
electronics housing to formerly be rather
chaotic. The application of a 3-D printed
ESCs holder has made it possible for all our ESCs Control
internal components to be put in their place Holder Board
in an orderly manner fig (7). This holder
resulted in improved cable management,
simpler maintenance and troubleshooting,
and shorter installation and disassembly.
PD
B
Anti-leakage system
An inventive anti-leakage system was
proposed and implemented employing Converter
raindrop sensors due to a problem that Interface
previously occurred and jeopardized our DRQ Board
sealing system and, consequently, our Converters
hardware system. This new system is
connected to a latching relay that shutdown
Heat Sink
the converters in the unfortunate event of a
leak, boosting our system's overall
Fig7. Onboard Electronics
efficiency and robustness in the process.
Image processing
Machine learning
Following the deployment of the appropriate filters to our live-streamed video feed. We employed optical
character recognition (OCR) using Python's OpenCV library for the zone number recognition task, which
allowed our system to execute advanced image recognition for the characters in the tasks with minimal
latency.
Measuring Objects
We developed an innovative method that can identify the length of any item by employing two cross lasers
spaced by a predetermined distance to compute an object's dimensions in real-time.