Surface Classification from Robot Internal Measurement Unit Time-Series Data Using Cascaded and Parallel Deep Learning Fusion Models
Machines,
Год журнала:
2025,
Номер
13(3), С. 251 - 251
Опубликована: Март 20, 2025
Surface
classification
is
critical
for
ground
robots
operating
in
diverse
environments,
as
it
improves
mobility,
stability,
and
adaptability.
This
study
introduces
IMU-based
deep
learning
models
surface
a
low-cost
alternative
to
computer
vision
systems.
Two
feature
fusion
were
introduced
classify
the
type
using
time-series
data
from
an
IMU
sensor
mounted
on
robot.
The
first
model,
cascaded
employs
1-D
Convolutional
Neural
Network
(CNN)
followed
by
Long
Short-Term
Memory
(LSTM)
network
then
multi-head
attention
mechanism.
second
model
parallel
which
processes
through
both
CNN
LSTM
simultaneously
before
concatenating
resulting
vectors
passing
them
Both
utilize
mechanism
enhance
focus
relevant
segments
of
time-sequence
data.
trained
normalized
Internal
Measurement
Unit
(IMU)
dataset,
with
hyperparameter
tuning
achieved
via
grid
search
optimal
performance.
Results
showed
that
higher
accuracy
metrics,
including
mean
Average
Precision
(mAP)
0.721
compared
0.693
model.
However,
incurred
44.37%
increase
processing
time,
makes
more
suitable
real-time
applications.
contributed
significantly
improvements,
particularly
Язык: Английский
Teaching Artificial Intelligence and Machine Learning in Secondary Education: A Robotics-Based Approach
Applied Sciences,
Год журнала:
2025,
Номер
15(8), С. 4570 - 4570
Опубликована: Апрель 21, 2025
The
rapid
advancement
of
Artificial
Intelligence
(AI)
and
Machine
Learning
(ML)
highlights
the
need
for
innovative,
engaging
educational
approaches
in
secondary
education.
This
study
presents
design
classroom
implementation
a
robotics-based
lesson
aimed
at
introducing
core
AI
ML
concepts
to
ninth-grade
students
without
prior
programming
experience.
intervention
employed
two
low-cost,
3D-printed
robots,
each
used
illustrate
different
aspect
intelligent
behavior:
(1)
rule-based
automation,
(2)
supervised
learning
using
image
classification,
(3)
reinforcement
learning.
was
compared
with
previous
similar
content
delivered
through
software-only
activities.
Data
were
collected
observation
student–teacher
discussions.
results
indicated
increased
student
engagement
enthusiasm
version,
as
well
improved
conceptual
understanding.
approach
required
no
specialized
hardware
or
instructor
expertise,
making
it
easily
adaptable
broader
use
school
settings.
Язык: Английский
Ensuring Safety in Human-Robot Cooperation: Key Issues and Future Challenges
Control Systems and Optimization Letters,
Год журнала:
2024,
Номер
2(3), С. 274 - 284
Опубликована: Ноя. 25, 2024
Human-robot
cooperation
(HRC)
is
becoming
increasingly
essential
in
many
different
sectors
such
as
industry,
healthcare,
agriculture,
and
education.
This
between
robot
human
has
advantages
increasing
boosting
productivity
efficiency,
executing
the
task
easily,
effectively,
a
fast
time,
minimizing
efforts
time.
Therefore,
ensuring
safety
issues
during
this
are
critical
must
be
considered
to
avoid
or
minimize
any
risk
danger
whether
for
robot,
human,
environment.
Risks
may
accidents
system
failures.
In
paper,
an
overview
of
human-robot
discussed.
The
main
key
challenges
robotics
outlined
presented
collision
detection
avoidance,
adapting
unpredictable
behaviors,
implementing
effective
mitigation
strategies.
difference
industrial
robots
cobots
illustrated.
Their
features
also
provided.
problem
avoidance
environment
defined
discussed
detail.
result
paper
can
guideline
framework
future
researchers
design
development
their
methods
tasks.
addition,
it
shapes
research
directions
measures.
Язык: Английский