Page 7 - IOE_Winter 24

AI-powered Car tested on Ice Rink provided us with a technologically sophisticated
Researchers at the University of Surrey have environment to operate in. The expertise
recently conducted an innovative experiment available, both in technology and business
at the Guildford Spectrum Ice Rink, which tested development, has been invaluable as we move
the University’s ZEBRA (Zero Emission test Bed towards the commercialisation phase of our
for Research on Autonomous driving) car with
business ”.
the objective of being able to use AI to develop a
reliable method to prevent skidding on wet or icy [ Reference: West, P., Electronic Specifier, 21st.
roads. November 2023 ]
The AI-equipped car was driven around the ice rink
for 90 minutes, providing a unique opportunity to test Reshaping 3D Electronics for AI Hardware
traction control in slippery conditions. The AI system
in the car is designed to detect when the wheels are
Collaborative research at the McKelvey School of
spinning too rapidly and adjust the power to prevent
Engineering at the Washington University in St.
skidding. The driverless function of the car was also
instrumental in allowing the researchers to control Louis has led to the successful demonstration of
the vehicle’s acceleration with greater accuracy, monolithic integration of layered 2D materials in
leading to more precise and consistent testing novel processing hardware for AI computing.
conditions.

The test led to the generation of more data than Multifunctional computer chips, known for their
anticipated, with the potential impact of the research integrated sensors, processor memory and
being described as ‘vast’ for enhancing road safety
specialised components, have traditionally
in adverse weather conditions.
expanded laterally, leading to increased time for
Further tests are planned at the rink later in 2024.
information transfer between components. This new
[ Reference: West, P., Electronic Specifier, 29th. development marks a paradigm shift away from
November 2023] this lateral integration by utilising six atomically thin
2D layers each of which fulfils a different function.
AI and Robotics help repair Potholes
Densely packed processing layers ensure robust
interlayer connectivity resulting in unparalleled
Around two million potholes are repaired annually
efficiency and performance in AI computing tasks.
on Britain’s roads, but traditional repair methods
are labour intensive and struggling to address the
growing problem. The breakthrough is seen as offering a new
approach for integrating electronics and the
Now, a UK start-up company known as Robotiz3d
has developed an autonomous robot that utilises AI dawn of ‘ a new era ‘ in multifunctional computing
to detect and repair cracks and potholes. hardware, the key lying in the ‘ultimate parallelism’
that could potentially vastly enhance the capabilities
Robotiz3d is located at the Science and Facilities
Council’s Daresbury Laboratory and the technology of AI systems.
originates from patented research at the University
of Liverpool that enables pothole geometry to be The leader of the research team, Assistant
analysed and measurement data to be collected.
Professor Sang Hoon Bae of the McKelvey School of
By combining this data with a unique Ai-driven
Engineering, states:
prediction algorithm local authorities will be able to
forecast road conditions and prioritise preventive
maintenance more accurately. It is also expected “ Monolithic 3D integration has the potential to
to reduce the time, cost, carbon emissions and reshape the entire electronics and computing
material waste associated with road repairs. industry by enabling the development of more
Lisa Layzell, CEO and Co-founder of Robotiz3d, compact, powerful and energy-efficient devices.
comments: Atomically thin 2D materials are ideal for this and
“ This is the first autonomous technology of this we are dedicated to refining this material until
kind developed specifically to tackle the pothole we can integrate all functional layers on a single
crisis which plagues many parts of the country, chip”.
and which is estimated to have cost more than £1
billion to repair over the last decade. For Robotiz3d [ Reference: West, P., Electronic Specifier, 10th.
locating to SFTC’s Daresbury Laboratory has December 2023]

7