Applied Artificial Intelligence,
Journal Year:
2024,
Volume and Issue:
39(1)
Published: Dec. 17, 2024
The
secure
transmission
of
communication
data
between
different
devices
still
faces
numerous
potential
challenges,
such
as
tampering,
integrity,
network
attacks,
and
the
risks
information
leakage
or
forgery.
This
approach
aims
to
handle
distributed
trust
issues
federated
learning
users
update
states
rapidly.
By
modeling
multi-source
through
learning,
model
parameters
reputation
values
participating
are
stored
on
blockchain.
method
incorporates
factors
experience,
familiarity,
timeliness
more
quickly
gather
reliable
about
nodes
assess
their
behavior.
Simulation
results
MNIST
dataset
show
that
when
proportion
selfish
is
below
50%,
convergence
time
increases
with
nodes.
Compared
advanced
algorithms,
proposed
saves
approximately
6%
interaction
time.
As
number
transactions
significantly
increases,
system's
TPS
(Transactions
Per
Second)
decreases,
an
average
only
3079.35
maximum
4000.
scheme
can
filter
out
high-quality
sources
during
real-time
dynamic
exchange,
enhancing
accuracy
training
ensuring
privacy
security.
Journal of Marine Science and Engineering,
Journal Year:
2024,
Volume and Issue:
12(8), P. 1371 - 1371
Published: Aug. 11, 2024
Due
to
limited
communication,
computing
resources,
and
unstable
environments,
traditional
cold
chain
traceability
systems
are
difficult
apply
directly
marine
scenarios.
Motivated
by
these
challenges,
we
construct
an
improved
blockchain-based
system
for
fishery
vessels.
Firstly,
Internet
of
Vessels
based
on
the
Iridium
Satellites
(IoV-IMS)
is
proposed
monitoring.
Aiming
at
problems
low
throughput,
long
transaction
latency,
high
communication
overhead
in
systems,
Practical
Byzantine
Fault
Tolerance
(PBFT)
consensus
algorithm,
a
Node-grouped
Reputation-evaluated
PBFT
(NR-PBFT)
improve
reliability
robustness
blockchain
system.
In
NR-PBFT,
node
grouping
scheme
designed,
which
introduces
consistent
hashing
algorithm
divide
nodes
into
candidate
sets,
reducing
number
participating
process,
lower
latency.
Then,
reputation
evaluation
model
selection
mechanism
NR-PBFT.
It
enhances
enthusiasm
participate
consensus,
considers
distance
between
vessels,
data
size,
refrigeration
temperature
factors
increase
throughput.
Finally,
carried
out
experiments
effectiveness
NR-PBFT
were
verified.
Compared
with
PBFT,
latency
shortened
81.92%,
throughput
increased
84.21%,
decreased
89.4%.
Applied Artificial Intelligence,
Journal Year:
2024,
Volume and Issue:
39(1)
Published: Dec. 17, 2024
The
secure
transmission
of
communication
data
between
different
devices
still
faces
numerous
potential
challenges,
such
as
tampering,
integrity,
network
attacks,
and
the
risks
information
leakage
or
forgery.
This
approach
aims
to
handle
distributed
trust
issues
federated
learning
users
update
states
rapidly.
By
modeling
multi-source
through
learning,
model
parameters
reputation
values
participating
are
stored
on
blockchain.
method
incorporates
factors
experience,
familiarity,
timeliness
more
quickly
gather
reliable
about
nodes
assess
their
behavior.
Simulation
results
MNIST
dataset
show
that
when
proportion
selfish
is
below
50%,
convergence
time
increases
with
nodes.
Compared
advanced
algorithms,
proposed
saves
approximately
6%
interaction
time.
As
number
transactions
significantly
increases,
system's
TPS
(Transactions
Per
Second)
decreases,
an
average
only
3079.35
maximum
4000.
scheme
can
filter
out
high-quality
sources
during
real-time
dynamic
exchange,
enhancing
accuracy
training
ensuring
privacy
security.
