IEEE Transactions on Consumer Electronics,
Journal Year:
2023,
Volume and Issue:
70(1), P. 1519 - 1530
Published: Dec. 5, 2023
With
the
emergence
of
Industry
5.0,
there
has
been
a
significant
surge
in
need
for
intelligent
services
within
realm
smart
devices.
Currently,
deep
neural
networks
(DNNs)
have
become
predominant
technology
driving
advancements
applications.
collaboration
mobile
edge
computing
(MEC),
resource-constraint
devices,
such
as
industrial
Internet
Things
(IIoT)
can
meet
requirement
high
DNN-based
inference
by
computation
offloading.
In
task
offloading
strategy
obtained
central
decision-maker
with
global
information,
all
devices
MEC
get
optimal
optimization
DNN
acceleration.
However,
practical
environment,
decision-making
may
into
trouble,
information
synchronization
delay,
irrational
behavior
and
privacy
leakage.
this
paper,
we
explore
distributed
to
deal
these
challenges
regarding
acceleration,
considering
character
an
early
exit
model
balance
accuracy
latency.
our
system
model,
is
formulated
decentralized
partially
observable
Markov
decision
process
(Dec-POMDP).
Each
device
performs
its
strategy,
including
branch
selection
local
observation,
cooperatively
optimizes
overall
Quality
Experience
inference.
Based
on
Dec-POMDP,
propose
one
algorithm
based
Multi-agent
Reinforcement
Learning
solve
above
problem.
algorithm,
utilize
advanced
function
counterfactual
baseline
guide
policy
gradient
learning
overcome
credit
allocation
problem
cooperative
optimization.
addition,
LSTM
introduced
improve
robustness
algorithm.
Finally,
detailed
performance
evaluation
comparison
are
performed
show
effectiveness
strategy.
International Journal of Remote Sensing,
Journal Year:
2025,
Volume and Issue:
unknown, P. 1 - 22
Published: Jan. 17, 2025
With
the
evolution
of
internet-of-things,
there
is
an
unprecedented
development
intelligent
sensors
owing
to
their
real-time
data
gathering
and
transmitting
capabilities.
A
number
these
are
deployed
in
remote
or
inaccessible
environment
harness
information
about
environmental
monitoring,
disaster
management,
climate
change,
wildlife
conservation,
marine
pollution,
natural
resource
management
precision
agriculture.
These
required
be
wirelessly
connected
provide
comprehensive
situational
details
share
abundant
data.
Thus,
needed
integrated
with
communication
systems
enable
efficient
decision
making
control.
This
paper
presents
sensing
(ISAC)
system
aided
reflecting
surfaces
(IRSs)
for
systems.
transmission
protocol
framed
assist
communication.
To
control
reduced
overhead,
a
beam
training
algorithm
proposed
that
aims
associate
beams
user
nodes
based
on
maximum
received
signal.
Further,
location
performed
utilizing
effective
angles-of-arrivals
information.
The
impact
transmit
power
Pt,
IRS-user
distance
dI2,k,
passive
elements
N1
count
N2
average
achievable
rate
localization
error
evaluated.
It
observed
method
achieves
13
bits/s/Hz
at
45
dBm
=
100.
improves
by
8%
scheme
when
sub-IRS
1
doubled.
Also,
10−3
obtained
10
64
dI2,k
m.
performance
comparison
conventional
random
also
investigated.
In
end,
integration
IRS-aided
ISAC
self-supervised
learning
analysing
plant
disease
detection
discussed
as
use
case
scenario.
2022 International Conference on Communication, Computing and Internet of Things (IC3IoT),
Journal Year:
2024,
Volume and Issue:
unknown, P. 1 - 6
Published: April 17, 2024
The
development
of
sophisticated
monitoring
systems
that
can
do
thorough
and
real-time
assessments
has
been
spurred
by
growing
worries
about
the
quality
water.
In
this
study,
we
suggest
a
unique
method
for
dynamically
water
combining
machine
learning
techniques
with
an
Internet
Things
(IoT)
sensor
network.
With
carefully
placed
IoT
sensors
inside
bodies
or
distribution
networks,
system
is
intended
to
continually
gather
multiple
parameter
data,
such
as
pH,
turbidity,
temperature,
dissolved
oxygen.
Modern
algorithms
housed
on
cloud
infrastructure
are
used
process
analyze
gathered
data.
Our
seeks
identify
abnormalities,
forecast
changes
in
quality,
offer
current
information
state
resources.
Machine
models
trained
past
data
order
detect
trends,
spot
departures
from
norm,
make
it
easier
proactive
decisions
reaction
possible
pollutants.
We
outline
design
our
network,
how
computing
integrated
processing,
put
into
practice
predictive
analytics.
also
go
over
system's
flexibility
changing
environmental
circumstances,
scalability,
uses
protection
resource
management.
Journal of Intelligent Manufacturing,
Journal Year:
2025,
Volume and Issue:
unknown
Published: Feb. 15, 2025
Abstract
Data
management,
particularly
in
industrial
environments,
is
increasingly
vital
due
to
the
necessity
of
handling
ever-growing
volumes
information,
commonly
referred
as
big
data.
This
survey
delves
into
various
papers
comprehend
practices
employed
within
settings
concerning
data
by
searching
for
relevant
keywords
Q1
Journals
related
management
manufacturing
databases
WebOfScience,
Scopus
and
IEEE.
Additionally,
a
contextual
overview
core
concepts
methods
different
aspects
process
was
conducted.
The
results
indicate
deficiency
methodology
across
implementations
even
same
types
industry
or
processes.
findings
also
highlight
several
key
principles
essential
constructing
an
efficient
optimized
system.
Journal of Quality in Maintenance Engineering,
Journal Year:
2025,
Volume and Issue:
unknown
Published: April 9, 2025
Purpose
In
Industry
4.0,
different
technologies
are
used
to
improve
the
efficiency
and
reduce
downtime
of
processes
in
organization.
It
can
be
achieved
by
using
predictive
maintenance
(PdM)
technique
avoid
sudden
breakdowns
industry.
is
important
implement
digital
twin
(DT)
for
PdM.
DT
PdM
nascent
stage.
This
study
focused
on
identification
determinants
real-life
implementation.
Design/methodology/approach
has
DTs
To
analyse
these
determinants,
multi-criteria
decision-making
(MCDM)
techniques
were
applying
Decision-Making
Trail
Evaluation
Laboratory
(DEMATEL)
interpretive
structural
modelling
(ISM)
approaches.
Findings
this
study,
13
found
out
through
literature
survey.
These
classified
into
cause
effect
DEMATEL
approach.
Similarly,
ISM
methodology
was
applied
categorized
levels.
results
compared,
it
that
real-time
analysis,
decision-making,
self-monitoring
diagnosis
most
important.
Practical
implications
useful
academic
researcher
as
well
industrialist
Therefore,
implemented
application
considering
determinants.
Originality/value
one
first
studies
represent
investigation
