Page 14 - Autumn 2024

Introducing Rhagobot the Water-skimming
Robot
Researchers at the University of California, Berkeley,
Ajou University in South Korea and the Georgia
Institute of Technology have developed a robot
called Rhagobot that mimics the movement of
Rhagovella, a small aquatic insect with fan-like feet
that can navigate turbulent conditions with a high
level of precision.
The feet of the Rhagovella unfold into fans when
submerged, providing ‘oars’ that let the insect glide
across water. No muscles are required as surface
tension provides the means of propulsion. Once lifted
from the water, the fans collapse, reducing drag and
preparing the insect for the next stroke.
Engineers at Ajou University created artificial versions
of the fans and attached them to the legs of an
insect-sized robot. Electron microscopy showed that
the natural fan is composed of flexible ribbon-like
strips equipped with barbules, a little like feathers.
When expanded underwater these barbules become
stiff, allowing them to act as oars.
Je-sung Koh, Professor at Ajou University and Senior
Author of the study, explains:
“Our robotic fans self-morph using nothing but
water surface forces and flexible geometry, just
like their biological counterparts. It is a form of
mechanical embedded intelligence refined by
nature through millions of years of evolution. In
small-scale robotics, these kinds of efficient and
unique mechanisms would be a key enabling
technology for overcoming limits in miniaturisation
of conventional robots”.
Co-author Saad Bhamla, Professor at Georgia Tech,
adds:
“We learned a rule from nature: the air-water
surface can act as a battery. Surface tension
powers the insect’s collapsible fan, and the same
design powers the robot fan”.
Rhagobot’s fans measure about 10x5 mm and are
mounted on the ends of two 5cm-long robotic legs.
It weighs 0.2 grams, can travel at around two body
lengths per second and turn 90 degrees in less than
half a second.
Reference: Hookway, P., ‘Tiny Insect inspires agile Water-skimming
Robot’, Electronic Specifier, 28th. October
14
E-skin provides Robots with Human-like Touch
Robotic interaction is poised to be transformed
following the development of an advanced next-
generation electronic skin for robots.
Postdoctoral research associate in electronic skin
Zhuo Chen, of Cambridge University, has used the
optimisation of polymer composite formulations and
advanced soft electronics fabrication to achieve spatial
resolution that surpasses human skin and sensitivity.
Robot bodies and dexterity thus far have tended
not to keep pace with the breakthroughs in AI that
are transforming robotic abilities. Most conventional
tactile sensors tend to be bulky, have low resolution
and are limited to detecting normal forces. The
capture of complex tactile information such as
the direction and magnitude of forces with high
resolution and accuracy has proved elusive. AI,
however, is now allowing for the translation of raw
sensor signals into meaningful tactile information.
It is enabling learning-based adaptation that
allows robots to refine their tactile capabilities in
dynamic environments where surface properties
are constantly changing. It also permits the e-skin
to adapt to different locations on a robot and meet
varied functional requirements.
The e-skin produces rich, multi-dimensional data,
including pressure magnitude, direction and
temperature which need to be rapidly interpreted for
decision-making.
Potential applications include healthcare, where it
can enhance the safety and precision of surgical
tools by providing real-time force feedback, and in
manufacturing where robots will be able to sense
contact forces, surface texture, stiffness and friction
with the ability to adjust their actions in real-time.
Zhuo Chen explains:
“Our e-skin uses a specially designed multiscale
composite architecture that decouples forces in
three dimensions. This enables accurate detection
of normal and shear forces, allowing robots
to sense sliding, texture, surface stiffness and
temperature with greater precision.
We’re exploring spiking neural networks and other
neuromorphic approaches that mimic the way the
human brain processes touch. This allows low-
latency, energy-efficient processing suitable for
embedded robotic systems”.
Reference: Ford, J., ‘E-skin to give Human-like Touch to Robots’, The
Engineer, September 2025, p.6
Become a member: https://membermojo.co.uk/ie
Become a subscriber: https://membermojo.co.uk/subscriber
Become a sponsor: https://membermojo.co.uk/sponsor