Frontiers in Neurorobotics,
Год журнала:
2023,
Номер
17
Опубликована: Ноя. 9, 2023
Deployment
of
Reinforcement
Learning
(RL)
algorithms
for
robotics
applications
in
the
real
world
requires
ensuring
safety
robot
and
its
environment.
Safe
Robot
RL
(SRRL)
is
a
crucial
step
toward
achieving
human-robot
coexistence.
In
this
paper,
we
envision
human-centered
SRRL
framework
consisting
three
stages:
safe
exploration,
value
alignment,
collaboration.
We
examine
research
gaps
these
areas
propose
to
leverage
interactive
behaviors
SRRL.
Interactive
enable
bi-directional
information
transfer
between
humans
robots,
such
as
conversational
ChatGPT.
argue
that
need
further
attention
from
community.
discuss
four
open
challenges
related
robustness,
efficiency,
transparency,
adaptability
with
behaviors.
While
interpretable
decision-making
is
pivotal
in
au-tonomous
driving,
research
integrating
natural
language
models
remains
a
relatively
untapped.
To
address
this,
we
introduce
multi-modal
instruction
tuning
dataset
that
facilitates
learning
visual
instructions
across
diverse
driving
scenarios.
This
encompasses
three
primary
tasks:
conversation,
detailed
description,
and
complex
reasoning.
Capitalizing
on
this
dataset,
present
LLM
assistant
named
VLAAD.
After
fine-tuned
from
our
instruction-following
VLAAD
demonstrates
proficient
interpretive
capabilities
spectrum
of
situations.
We
open
work,
model,
to
public
github.
https://github.
com/sungyeonparkk/vision-assistant-for-driving
Multimodal Technologies and Interaction,
Год журнала:
2025,
Номер
9(1), С. 6 - 6
Опубликована: Янв. 14, 2025
Multimodal
interaction
is
a
transformative
human-computer
(HCI)
approach
that
allows
users
to
interact
with
systems
through
various
communication
channels
such
as
speech,
gesture,
touch,
and
gaze.
With
advancements
in
sensor
technology
machine
learning
(ML),
multimodal
are
becoming
increasingly
important
applications,
including
virtual
assistants,
intelligent
environments,
healthcare,
accessibility
technologies.
This
survey
concisely
overviews
recent
interaction,
interfaces,
communication.
It
delves
into
integrating
different
input
output
modalities,
focusing
on
critical
technologies
essential
considerations
fusion,
temporal
synchronization
decision-level
integration.
Furthermore,
the
explores
challenges
of
developing
context-aware,
adaptive
provide
seamless
intuitive
user
experiences.
Lastly,
by
examining
current
methodologies
trends,
this
study
underscores
potential
sheds
light
future
research
directions.
International Journal of Innovation Science,
Год журнала:
2025,
Номер
unknown
Опубликована: Фев. 10, 2025
Purpose
This
study
aims
to
investigate
the
evolution
of
cybersecurity
in
autonomous
vehicles
over
past
decade,
focusing
on
influential
publications,
leading
authors,
key
themes
and
emerging
research
trends.
Design/methodology/approach
A
systematic
literature
review
was
conducted
using
Preferred
Reporting
Items
for
Systematic
Reviews
Meta-Analyses
approach,
with
data
extracted
from
The
Lens
database
analyzed
VOSviewer
Bibliometrix.
provides
a
quantitative
overview
academic
trends
2014
2023.
analysis
reveals
significant
growth
scientific
production,
predominantly
driven
by
USA,
China
UK.
Central
include
network
security,
cyberattack
prevention
regulatory
frameworks.
Findings
findings
emphasize
that
cybersecurity,
artificial
intelligence
(AI)
regulation
are
critical
developing
secure
reliable
vehicular
systems.
Research
limitations/implications
Future
should
focus
enhancing
security
vehicle-to-everything,
vehicle-to-vehicle
vehicle-to-infrastructure
communications
improving
protocols
integrating
AI.
Practical
implications
Key
identified
trust
reliability
user
experience.
Social
highlights
future
directions,
particularly
integration
AI
sustainable
development
transportation
policies.
Originality/value
2023
regarding
theme
self-driving
cars.
Systems Research and Behavioral Science,
Год журнала:
2025,
Номер
unknown
Опубликована: Фев. 22, 2025
ABSTRACT
Background:
The
traditional
human–vehicle
relationship
and
the
challenges
posed
by
complex
driving
scenarios
have
led
to
situations
where
drivers
experience
‘Out
of
Loop’
(OOTL)
cognition,
resulting
in
inefficient
communication
a
threat
safety.
Purpose:
cognitive
state
an
interactive
environment
significantly
influences
level
collaborative
efficiency.
This
study
investigates
logic
interaction
modes
intelligent
systems
that
promote
driver
cognition
loop,
aiming
improve
Methods:
paper
addresses
issue
loop
within
collaboration
through
knowledge
graphs
literature
reviews
elucidate
evolution
relationships
analyse
key
elements
collaboration.
By
examining
characteristics
behaviours
during
driver's
perception,
understanding,
prediction,
decision‐making
action
phases,
it
summarizes
impact
mechanisms
solutions
understanding
prediction
as
well
on
tasks.
Finally,
provides
design
strategies
evaluation
methods
for
development
cockpit
design.
