IEEE Transactions on Consumer Electronics,
Год журнала:
2023,
Номер
70(2), С. 4922 - 4933
Опубликована: Ноя. 14, 2023
As
one
of
the
consumer
electronics,
drones
have
received
more
and
attention
in
applications
such
as
dynamic
monitoring,
transportation
goods,
unmanned
logistics.
In
these
applications,
it
is
necessary
to
obtain
accurate
position
drone
can
track
navigate.
The
usually
work
outdoors,
GPS
signals
are
most
important
information
their
position.
Therefore,
GPS-based
positioning
a
key
research
issue
for
applications.
On
other
hand,
this
puts
forward
higher
requirements
mobile
target
tracking
technology.
tracker
real-time
dynamics
location
through
state
estimation.
Classical
estimation
methods
require
system
model
with
Gaussian
white
noise,
whereas
data
contains
pink
making
an
exact
match
actual
challenging.
This
study
uses
new
end-to-end
implemented
data-driven
mechanism
incorporating
codec
catch
complex
nonlinear
dynamics.
Further,
switchable
fuzzy
normalization
loaded
codec,
using
three
different
algorithms,
z-score,
adaptively
process
input
data,
realize
measurement
data's
adaptive
correction
extract
features
GPS.
We
experimentally
conclude
that
compared
classical
deep
learning
model,
method
proposed
paper
reduces
RMSE
by
average
7.95%
20.6%,
respectively,
optimal
avoids
modelling
system,
efficiently
overcome
chromatic
noise
observation,
outperforms
trajectory
performance
filtering
methods.
Agriculture,
Год журнала:
2023,
Номер
13(3), С. 567 - 567
Опубликована: Фев. 26, 2023
In
modern
agriculture
and
environmental
protection,
effective
identification
of
crop
diseases
pests
is
very
important
for
intelligent
management
systems
mobile
computing
application.
However,
the
existing
mainly
relies
on
machine
learning
deep
networks
to
carry
out
coarse-grained
classification
large-scale
parameters
complex
structure
fitting,
which
lacks
ability
in
identifying
fine-grained
features
inherent
correlation
mine
pests.
To
solve
problems,
a
pest
method
based
graph
pyramid
attention,
convolutional
neural
network
(GPA-Net)
proposed
promote
agricultural
production
efficiency.
Firstly,
CSP
backbone
constructed
obtain
rich
feature
maps.
Then,
cross-stage
trilinear
attention
module
extract
abundant
discrimination
portions
objects
as
much
possible.
Moreover,
multilevel
designed
learn
multiscale
spatial
graphic
relations
enhance
recognize
diseases.
Finally,
comparative
experiments
executed
cassava
leaf,
AI
Challenger,
IP102
datasets
demonstrates
that
GPA-Net
achieves
better
performance
than
models,
with
accuracy
up
99.0%,
97.0%,
56.9%,
respectively,
more
conducive
distinguish
applications
practical
smart
protection.
IEEE Transactions on Intelligent Transportation Systems,
Год журнала:
2023,
Номер
25(3), С. 2966 - 2975
Опубликована: Май 24, 2023
Accurate
traffic
flow
prediction,
a
hotspot
for
intelligent
transportation
research,
is
the
prerequisite
prediction
making
travel
plans.
The
speed
of
can
be
affected
by
roads
condition,
weather,
holidays,
etc.
Moreover,
sensors
to
catch
information
about
will
interfered
with
environmental
factors
such
as
illumination,
collection
time,
occlusion,
Therefore,
in
practical
system
complicated,
uncertain,
and
challenging
predict
accurately.
Motivated
from
aforementioned
issues
challenges,
this
paper,
we
propose
deep
encoder-decoder
framework
based
on
variational
Bayesian
inference.
A
neural
network
designed
combining
inference
Gated
Recurrent
Units
(GRU)
which
used
unit
mine
intrinsic
dynamics
flow.
Then,
introduced
into
multi-head
attention
mechanism
avoid
noise-induced
deterioration
accuracy.
proposed
model
achieves
superior
performance
Guangzhou
urban
dataset
over
benchmarks,
particularly
when
long-term
prediction.
IECE transactions on intelligent systematics.,
Год журнала:
2024,
Номер
1(2), С. 79 - 90
Опубликована: Сен. 27, 2024
Accurate
predictions
of
traffic
flow
are
very
meaningful
to
city
managers.
With
such
information,
systems
can
better
coordinate
signals
and
reduce
congestion.
By
understanding
patterns,
navigation
provide
real-time
routing
suggestions
that
avoid
jams,
save
time,
fuel
consumption.
However,
will
be
interfered
with
by
multiple
factors
as
collection
time
place.
In
this
paper,
we
propose
use
stochastic
configuration
networks
(SCNs)
predict
the
flow.
The
network
is
trained
through
stepwise
construction,
parameters
effectively
optimized
based
on
approximation
theorem
convergence
analysis
optimization
mechanism.
proposed
automatically
adjusts
its
structure
according
complexity
adapt
complex
non-linearity
We
observed
model
achieves
prediction
performance
overall
greater
flexibility
in
length
period
compared
benchmarks
using
Guangzhou
urban
dataset.
It's
worth
noting
SCNs
consistently
outperform
other
models
across
different
intervals.
They
yield
RMSE
improvements
up
10.73%
for
10-minute
predictions,
5.02%
30-minute
11.21%
60-minute
least
effective
models.
R-value
also
exhibits
steady
enhancement,
increasing
0.78%,
0.65%,
2.33%
10-minute,
30-minute,
respectively.
These
notable
advancements,
combined
model's
computational
efficiency,
especially
short-term
underscore
effectiveness
practicality
tasks.
Journal of Cloud Computing Advances Systems and Applications,
Год журнала:
2024,
Номер
13(1)
Опубликована: Март 21, 2024
Abstract
Powered
by
data-driven
technologies,
precision
agriculture
offers
immense
productivity
and
sustainability
benefits.
However,
fragmentation
across
farmlands
necessitates
distributed
transparent
automation.
We
developed
an
edge
computing
framework
complemented
auction
mechanisms
fuzzy
optimizers
that
connect
various
supply
chain
stages.
Specifically,
powerful
capabilities
enable
real-time
monitoring
decision-making
in
smart
agriculture.
propose
tailored
to
agricultural
needs
ensure
through
a
renewable
solar
energy
supply.
Although
the
manages
crop
data
collection,
market-based
mechanisms,
such
as
auctions
optimization
models,
support
for
smooth
operations.
formulated
invisible
hide
actual
bid
values
regulate
information
flows,
combined
with
machine
learning
techniques
robust
predictive
analytics.
While
rule-based
systems
encode
domain
expertise
decision-making,
adaptable
training
algorithms
help
optimize
model
parameters
from
data.
A
two-phase
hybrid
approach
is
formulated.
Fuzzy
models
were
using
three
key
decision
problems.
Auction
markets
discover
optimal
demand–supply
balancing
pricing
signals.
incorporate
knowledge
into
interpretable
crop-advisory
models.
An
integrated
evaluation
of
50
farms
over
five
cycles
demonstrated
high
performance
proposed
computing-oriented
auction-based
neural
network
compared
benchmarks.
Energies,
Год журнала:
2024,
Номер
17(2), С. 415 - 415
Опубликована: Янв. 15, 2024
Load
forecasting
is
a
research
hotspot
in
academia;
the
context
of
new
power
systems,
prediction
and
determination
load
reserve
capacity
also
important.
In
order
to
adapt
forms
day-ahead
automatic
generation
control
(AGC)
demand
method
based
on
Fourier
transform
attention
mechanism
combined
with
bidirectional
long
short-term
memory
neural
network
model
(Attention-BiLSTM)
optimized
by
an
improved
whale
optimization
algorithm
(IWOA)
proposed.
Firstly,
response
time,
used
refine
distinction
between
various
types
demand,
AGC
band
calculated
using
Parseval’s
theorem
obtain
sequence.
The
maximum
mutual
information
coefficient
explore
relevant
influencing
factors
sequence
concerning
data
characteristics
Then,
historical
daily
sequences
features
are
input
into
Attention-BiLSTM
model,
automatically
find
optimal
hyperparameters
better
results.
Finally,
arithmetic
simulation
results
show
that
proposed
this
paper
has
best
performance
upper
(0.8810)
lower
(0.6651)
bounds
(R2)
higher
than
other
models,
it
smallest
mean
absolute
percentage
error
(MAPE)
root
square
(RMSE).
Applied Sciences,
Год журнала:
2025,
Номер
15(2), С. 912 - 912
Опубликована: Янв. 17, 2025
Mental
health
disorders
constitute
a
significant
global
challenge,
compounded
by
the
limitations
of
traditional
management
approaches
that
rely
heavily
on
subjective
self-reports
and
infrequent
professional
evaluations.
This
study
presents
groundbreaking
IoT-based
system
integrates
big
data
analytics,
fuzzy
logic,
machine
learning
to
revolutionise
mental
monitoring.
In
contrast
existing
solutions,
proposed
uniquely
incorporates
environmental
factors,
such
as
temperature
humidity
in
enclosed
spaces—critical
yet
often
overlooked
contributors
emotional
well-being.
By
leveraging
IoT
devices
collect
process
large-scale
ambient
data,
provides
real-time
classification
personalised
visualisation
tailored
individual
sensitivity
profiles.
Preliminary
results
reveal
high
accuracy,
scalability,
potential
generate
actionable
insights,
creating
dynamic
feedback
loops
for
continuous
improvement.
innovative
approach
bridges
gap
between
conditions
healthcare,
promoting
transformative
shift
from
reactive
proactive
care
laying
groundwork
predictive
systems.
Abstract
Automated
disease
recognition
plays
a
pivotal
role
in
advancing
smart
artificial
intelligence
(AI)‐based
agriculture
and
is
crucial
for
achieving
higher
crop
yields.
Although
substantial
research
has
been
conducted
on
deep
learning‐based
automated
plant
systems,
these
efforts
have
predominantly
focused
leaf
diseases
while
neglecting
affecting
fruits.
We
propose
an
efficient
architecture
effective
fruit
with
state‐of‐the‐art
performance
to
address
this
gap.
Our
method
integrates
advanced
techniques,
such
as
multi‐head
attention
mechanisms
lightweight
convolutions,
enhance
both
efficiency
performance.
Its
ultralightweight
design
emphasizes
minimizing
computational
costs,
ensuring
compatibility
memory‐constrained
edge
devices,
enhancing
accessibility
practical
usability.
Experimental
evaluations
were
three
diverse
datasets
containing
multi‐class
images
of
disease‐affected
healthy
samples
sugar
apple
(
Annona
squamosa
),
pomegranate
Punica
granatum
guava
Psidium
guajava
).
proposed
model
attained
exceptional
results
test
set
accuracies
weighted
precision,
recall,
f1‐scores
exceeding
99%,
which
also
outperformed
pretrain
large‐scale
models.
Combining
high
accuracy
represents
significant
step
forward
developing
accessible
AI
solutions
agriculture,
contributing
the
advancement
sustainable
agriculture.
Chinese journal of information fusion.,
Год журнала:
2025,
Номер
2(1), С. 38 - 58
Опубликована: Март 22, 2025
With
the
progressive
advancement
of
remote
sensing
image
technology,
its
application
in
agricultural
domain
is
becoming
increasingly
prevalent.
Both
cultivation
and
transportation
processes
can
greatly
benefit
from
utilizing
images
to
ensure
adequate
food
supply.
However,
such
often
exist
harsh
environments
with
many
gaps
dense
distribution,
which
poses
major
challenges
traditional
target
detection
methods.
The
frequent
missed
detections
inaccurate
bounding
boxes
severely
constrain
further
analysis
within
sector.
This
study
presents
an
enhanced
version
YOLO
algorithm,
specifically
tailored
achieve
high-efficiency
densely
distributed
small
targets
images.
We
replaced
convolutions
a
convolution
kernel
size
3
last
two
ELAN
modules
DeformableConvNetsv2
so
that
backbone
better
extract
various
objects.
proposed
detector
introduces
Bi-level
Routing
Attention
module
pooled
pyramid
SPPCSPC
network
YOLOv7,
thereby
intensifying
attention
towards
areas
concentration
augmenting
network's
capacity
features
related
through
effective
feature
fusion.
Additionally,
our
approach
employs
dynamic
non-monotonic
WIoUv3
loss
function
network,
enabling
allocation
most
appropriate
gradient
gain
strategy
at
each
instant
enhancing
ability
focus
on
detecting
accurately.
Finally,
comparative
experimentation
DIOR
dataset,
YOLOv7-bw
exhibits
superior
performance
higher
[email protected][email protected]:
0.95,
achieving
rates
85.63%
65.93%,
surpassing
those
YOLOv7
by
1.93%
2.03%,
respectively,
thus
substantiating
effectiveness
algorithmic
approach.