Page 9 - Autumn 2024
P. 9
Robotic Berry Harvester matches Human Pickers Breakthrough helps free up Space for AI Development for Wind Turbine Another aspect of the project is the coordination
Robots to ‘think’ Autonomous Robots of Remotely Operated Vehicles (ROVs) and Electric
A robotic berry harvester, known as Fieldworker
ROVs deployed from Uncrewed Surface Vessels,
1, can now harvest berries at the same speed Researchers at King’s College London The National Robotarium in Scotland is assisting
which is important for achieving fully autonomous
and to the same quality as human pickers, have devised a way to give robots complex the application of AI and control systems that
inspections and reducing the need for support
providing growers with a predictable cost instructions without the use of electricity for the could enable underwater robots to operate
vessels and divers.
per berry. first time, which is envisaged to free up more autonomously with high precision in turbulent
space in the robotic ‘brain’ to allow robots to areas, potentially revolutionising maintenance David Morrison, Project Manager at the National
Fieldworker 1 by Fieldwork Robotics is powered by
‘think’. and repair of offshore wind turbines. Robotarium, comments:
AI-based technology that allows it to detect berry
ripeness accurately with the aid of four global The technique involves transmitting a series Heriot-Watt University is currently leading the “ Our trials are showing promising results in
cameras, eight end effector cameras and of commands using a new type of compact project with the support of Imperial College enabling underwater robots to maintain stable
spectral frequency. circuit that utilises variations in pressure of a London, geo-data specialist Fugro and underwater contact with offshore structures in challenging
fluid inside it, creating the potential for a new software development expert Frontier Robotics. conditions. If successful, the technology could
generation of robots with bodies that could operate transform offshore wind maintenance, potentially
independently from their built-in control centre, with reducing fuel consumption of maintenance
this space being able to be deployed instead for missions by up to 97 per cent, from 7,000 litres per
more complex AI-powered software. day to just 200 litres. This could significantly lower
both operational costs and the carbon footprint of
The innovation is also envisaged to enable
maintenance.”
the creation of robots that can operate where
electrically powered devices cannot work, such Reference: Miles, S., ‘AI enables Autonomous Robots for precision
Offshore Wind Farm Repairs’, Electronic Specifier, 25th. October.
as irradiated areas that destroy circuits, and in
electrically sensitive environments such as MRI
rooms.
The technology involves a reconfigurable circuit
with an adjustable valve, which is incorporated into
the robot’s hardware. This valve acts in a manner
analogous to that of a transistor in a normal circuit
such that engineers can send signals directly to
hardware using pressure to mimic binary code. The
Each payload has four arms with each arm being
robot can then perform complex activities without
positioned on a z-axis so each arm can travel up
and down the z-axis for maximum height and
reach. Each z-axis has a dedicated global camera
that scans the bush to detect the location and
ripeness of the berries. When a suitable berry is
identified the arm moves up the z-axis to a position
such that it can extend out on an x-y movement. The objective is to dramatically reduce the
At the end of each arm two further cameras are requirement for large maintenance vessel
located on the end effector, which is in effect a cup deployment in offshore wind farm operations and
that removes the berry from the bush. for personnel to have to work in hazardous offshore
environments. Through this project autonomous
It is intended that fleets of Fieldworker 1 robots
underwater robots could perform many essential
will operate with one operator running multiple
maintenance tasks on offshore wind turbines,
robots across the field. Both day and night the need for electricity or instruction from a central
including taking precise measurements,
operation is envisaged and the first trials are set ‘brain’, so allowing for a greater level of control than
conducting visual inspections, cleaning structures
to be conducted on berry farms in Australia in the that which is possible with current fluid-based
and repairing identified defects.
coming months. circuits.
A key challenge in such deployments is keeping
Post-graduate Researcher at King’s College
Reference: Ford, J., The Engineer, October 2024, p.9. the robot arm or tool steady against a structure
London, and author of the study, Mostafa Monsa,
whilst being buffeted by currents and waves, and
forecasts:
to achieve this advanced control systems and
“ Ultimately, without investment in embodied machine learning algorithms are being developed
Visit our website! intelligence robots will plateau. Soon, if we do not to allow the robots to adapt to real-time changing
offload the computational load that modern day conditions. Additionally, 3D semantic mapping
Why not head down to our website
and read more news in our blog? robots take on, algorithmic improvements will have capabilities are being developed so that the robots
little impact on their performance. Our work is just a can create detailed maps of their underwater
Updated daily with hand-picked
articles from around the web... first step on this path, but the future holds smarter environments, with the robots subsequently
robots with smarter bodies.” enabled to navigate complex structures and
identify components that require attention.
institutionofelectronics.ac.uk Reference: Miles,S., Electronic Specifier, 11th. October
8 9
Subscribe to Electron here https://membermojo.co.uk/ie Subscribe to Electron here https://membermojo.co.uk/ie

