Page 12 - Winter 2025

Industry News
FURTHER ADVANCES
IN ROBOTICS
Harnessing Brain Function for Robot Navigation
One area of robotics that continues to pose
challenges is navigation. AI-derived navigation
systems are commonly used, but have significant
computational and energy requirements.
Now, a team at the Queensland University of
Technology has developed a novel place recognition
algorithm that uses Spiking Neural Networks, artificial
neural networks that mimic how biological brains
process information using brief, discrete signals.
These networks are noted to be particularly suitable
for neuromorphic hardware, that is to say computer
hardware that mimics biological neural systems
so as to enable faster processing and significantly
reduced energy consumption.
The system uses small neural network modules
to recognise specific places from images. These
modules were combined into an ensemble, that is a
group of multiple spiking networks, to form a scalable
navigation system that can learn to navigate in large
environments.
Dr. Tobias Fischer of the Queensland University of
Technology Centre for Robotics comments:
“ Animals are remarkably adept at navigating
large, dynamic environments with amazing
efficiency and robustness. This work is a step
towards the goal of biologically inspired navigation
systems that could one day compete with or even
surpass today’s more conventional approaches”.
12
Colleague Professor Michael Milford adds:
“ Using sequences of images instead of single
images enabled an improvement of 41 per cent in
place recognition accuracy, allowing the system
to adapt to appearance changes over time and
across different seasons and weather conditions”.
Leader of the team, postdoctoral research fellow
Somayeh Hussaini, concluded:
“ This work can help pave the way for more efficient
and reliable navigation systems for autonomous
robots in energy-constrained environments.
Particularly exciting opportunities include domains
like space exploration and disaster recovery,
where optimising energy efficiency and reducing
response times are critical”.
Reference: Gittins, C, ‘Research harnesses Brain Function for Robot
Navigation’, Electronic Specifier, 4th. December
Wearable Robot provides Parkinson’s Support
Scientists at the Harvard John A. Paulson School of
Engineering and Applied Sciences and at the Boston
Sargent College of Health and Rehabilitation Service
have successfully tested a soft, wearable robot that
allows a person with Parkinson’s to walk without
freezing.
Parkinson’s reportedly affects over 9 million people
worldwide and freezing is a common symptom.
When freezing occurs the patient loses the ability to
move their feet mid-stride, causing stuttering steps
that worsen until they eventually stop.
The robotic garment is worn around the hips and
thighs and provides a push to the hips as the leg
swings, so giving the person a longer stride.
Become a member: https://membermojo.co.uk/ie
Become a subscriber: https://membermojo.co.uk/subscriber
Become a sponsor: https://membermojo.co.uk/sponsor
The development followed a pilot with a 73-year-old
man who, despite receiving conventional treatment,
repeatedly froze up to ten times a day. Earlier
research used human-in-the-loop optimisation to
prove that a soft, wearable device could be used to
augment hip flexion and assist in swinging the leg
forward. This principle was extended such that it used
cable-driven actuators and sensors and by using
motion data algorithms were developed to estimate
the phase of the gait enabling assistive forces to be
generated in tandem with muscle movement.
Reference: Gittins, C., ‘Soft wearable Robot tested for Parkinson’s
Support’, Electronic Specifier, 4th. November
Agriculture and Robotics combine to improve
Soil Health
The University of Plymouth has joined with startup
robotics company Robotriks in a project to improve
soil health using a robotic dog.
Traditional soil testing usually requires labour-
intensive processes, such as collecting samples with
trowels, followed by laboratory analysis which can
take weeks of months to deliver results.
The new robotic dog uses advanced gamma-
ray detection technology to scan soil for naturally
occurring radiation levels to assess its overall
Become a member: https://membermojo.co.uk/ie
Become a subscriber: https://membermojo.co.uk/subscriber
Become a sponsor: https://membermojo.co.uk/sponsor
health. Moving at up to five metres per second it
can navigate challenging terrain and reach areas
that would typically be inaccessible for farmers and
conventional machinery and provide fast field-
wide data. Detailed maps of soil properties that are
relevant to farming decisions can be provided and
farmers can use the insights to manage land more
effectively, adapting to changes in soil conditions
as they occur rather than having to react weeks or
months later.
Reference: Miles, S., ‘Robotic Dog advises Farmers on healthy Soil’,
Electronic Specifier, 26th. November
Automating Order Picking with AI and Robotics
A supermarket in Germany that is situated close to
a frozen pizza manufacturer and needs a consistent
supply of pizzas has turned to artificial intelligence
to automatically pack the different types of pizzas.
Seventeen participants from two Fraunhofer
Institutes, the Lamarr Institute for Machine Learning
and Artificial Intelligence and employees from the
University of Bonn came together at a ‘hackathon’
hosted by igus in Cologne to develop the ingenious
solution.
Three interdisciplinary teams were challenged to
program a robot to recognise the right products
from voice instructions from employees, pick them
up and pack them into mixed packages. With the
aid of an igus ReBel robotic arm placed in front of a
conveyor belt on which were placed various pizzas,
a webcam and an AI-based segmentation system
(Segment Anything Model), the team developed a
system that could identify the various objects on the
conveyor belt and their position. A language model
mapped the objects to the right products based on
the language instructions, and placed the products
in the correct boxes by following the instructions of
the language model.
Alexander Zorn of the Fraunhofer Institute for
Intelligent Analysis and Information Systems, states:
13