Network Functional Compression for Control Applications Sifat Rezwan Juan A. Cabrera and Frank H. P. Fitzek

2025-05-02 0 0 4.96MB 5 页 10玖币
侵权投诉
Network
Functional
Compression
for
Control
Applications
Sifat
Rezwan*,
Juan
A.
Cabrera*,
and
Frank
H.
P.
Fitzek*!
*Deutsche
Telekom
Chair
of
Communication
Networks,
TU
Dresden,
Germany
t
Centre
for
Tactile
Internet
with
Human-in-the-Loop
(CeTI)
{sifat.rezwan
|
juan.cabrera
|
frank.fitzek}
@tu-dresden.de
Abstract—The
trend
of
future
communication
systems
is
to
aim
for
the
steering
and
control
of
cyber-physical
systems.
These
systems
can
quickly
become
congested
in
environments
like
those
presented
in
Industry
4.0.
In
these
scenarios,
a
plethora
of
sensor
data
is
transmitted
wirelessly
to
multiple
in-network
controllers
that
compute
the
control
functions
of
the
cyber-physical
systems.
In
this
paper,
we
show
an
implementation
of
network
Functional
Compression
(FC)
as
a
proof
of
concept
to
drastically
reduce
the
data
traffic
in
these
scenarios.
FC
is
a
form
of
goal-oriented
communication
scheme
in
which
the
objective
of
the
sender-
receiver
pair
is
to
transmit
the
minimum
amount
of
information
to
compute
a
function
at
the
receiver
end.
In
our
scenario,
the
senders
transmit
an
encoded
and
compressed
version
of
the
sensor
data
to
a
destination,
an
in-network
controller
interested
in
computing
as
its
target
function,
a
PID
controller.
We
show
that
it
is
possible
to
achieve
compression
rates
of
over
50%
in
some
cases
by
employing
FC.
We
also
show
that
using
FC
in
a
distributed
cascade
fashion
can
achieve
more
significant
compression
rates
while
reducing
computational
costs.
Index
Terms—Functional
compression,
goal-orientated
com-
munication,
graph
colouring,
in-network
computing,
post-
Shannon.
I.
INTRODUCTION
In
the
revolutionary
classical
information
theory,
Shannon
solely
focused
on
the
statistics
of the
messages
ignoring
the
semantics
and
the
syntax
to
lay
the
foundations
for
the
information
age
[1].
From
Shannon’s
point
of
view,
the
fundamental
communication
problem
is
replicating
messages
from
one
point
to
another.
This
perspective
completely
ignores
the
intended
goal
of the
replicated
messages.
However,
with
the
advancement
of
communication
networks,
the
quality
of
service
(QoS)
requirements,
such
as
reliability,
latency,
secu-
rity,
and
robustness,
are
getting
more
strict.
We
are
reaching
the
capacity
of
our
communication
systems,
and
therefore,
we
must
consider
approaches
that
go
beyond
the
traditional
goal-agnostic
message
replication
paradigm.
This
will
require
addressing
the
aforementioned
strict
QoS
policies
of
massive
and
heterogeneous
future
communication
systems.
These
ap-
proaches
are
beneficial
for the
low-latency
and
ultra-reliable
communication
systems
for
steering
and control
applications
of
the,
e.g.,
Industry
4.0.
This
type
of
study
can
be
referred
to
as
goal-oriented
or
post-Shannon
communication
[2],
[3].
The
main
idea
of
post-Shannon
communication
is
to
extend
Shannon’s
communication
framework
to
consider
the
seman-
tics
of
the
messages,
the
intended
goals
of
communication,
or
both
to
reduce
the
required
bandwidth
of
communication
978-1
-6654-7095-7/22/$31.00
©
2022
IEEE
This
work
has
been
accepted
for
publication
by
IEEE
at
the
International
Conference
on
Electrical,
Computer,
Communications
and
Mechatronics
Engineering
(ICECCME),
16-18
November
2022,
Maldives.
further.
It
is
primarily
applicable
to
the
scenarios
where
the
receivers
have
to
achieve
a
particular
goal
or
perform
a
specific
task.
Post-Shannon
theory
includes
different
approaches
to
address
different
types,
such
as
message
identification
[4],
[5],
common
randomness,
functional
compression
(FC)
[6],
and
medium
as
the
message
[7]
for
goal-oriented
use
cases.
In
this
paper,
we
solely
focus
on
FC
as
a
potential
post-
Shannon
technique
to
reduce
the
required
data
rates
for
in-
network
computing
applications
for
steering
and
control.
The
fundamental
problem
of
FC
is
how
much
information
the
sources
should
transmit
if
the
destination
is
interested
in
computing
a
function
of the
sources.
FC
can
be
seen
as
a
generalisation
of
Shannon’s
source
compression
problem.
In
the
traditional
Shannon
scenario,
the
primary purpose
of
data
compression
is
to
remove
the
redundancy
of the
source
to
reconstruct
it
as
accurately
as
possible
at
the
receiver
end.
In
a
sense,
it
answers
the
question
of
how
much
information
the
sources
should
transmit
to
compute
the
identity
function
(a
function
whose
output
is
the
same
as
its
inputs)
at
the
receiver.
For
example,
for the
distributed
case
of
source
compression,
the
Slepian-Wolf
theorem
describes
the
achievable
compres-
sion
rates
[8].
In
FC,
however,
the
problem
is
generalised
to
reconstruct
the
output
of
any
arbitrary
function
at
the
destination
instead
of
solely
focusing
on
regenerating
the
source
information
[9].
Therefore,
the
Slepian-Wolf
source
encoding
example
is
a
particular
case
of
FC
where
the
function
to
be
computed
at
the
receiver
end
is
the
identity
function
of
the
sources.
In
this
paper,
we
show,
as
a
proof-of-concept,
an
imple-
mentation
of
FC
for
in-network
control
applications
based
on
proportional—integral—derivative
(PID)
controllers.
In
our
scenario,
a
PID
controller
located
in
the
network
is
interested
in
computing
a
control
function
based
on
the
sensor
data
of
a
plant.
We
use
FC
to
compress
the
source’s
sensor
data
to
transmit
to
the
controller
to
calculate
the
control function,
thus
saving
network
resources.
The
remaining
of
this
paper
is
structured
as
follows.
In
Section
II,
we
introduce
the
fundamentals
of
FC.
In
Section
III,
we
describe
our
scenario
of
an
in-network
PID
controller
controlling
the
water
level
in
a
water
tank.
In
Section
IV
we
make
a
comparative
analysis
of
two
FC
approaches
for
solving
the
problem.
In
Section
V
we
discuss
some
open
issues
of
FC
for
practical
implementations,
and
finally,
in
Section
VI
we
present
our
conclusions.
摘要：

NetworkFunctionalCompressionforControlApplicationsSifatRezwan*,JuanA.Cabrera*,andFrankH.P.Fitzek*!*DeutscheTelekomChairofCommunicationNetworks,TUDresden,GermanytCentreforTactileInternetwithHuman-in-the-Loop(CeTI){sifat.rezwan|juan.cabrera|frank.fitzek}@tu-dresden.deAbstract—Thetrendoffuturecommunica...

展开>> 收起<<
Network Functional Compression for Control Applications Sifat Rezwan Juan A. Cabrera and Frank H. P. Fitzek.pdf

共5页,预览1页

还剩页未读， 继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

相关推荐

更多
立即下载
分类：图书资源 价格：10玖币 属性：5 页 大小：4.96MB 格式：PDF 时间：2025-05-02

开通VIP享超值会员特权

  • 多端同步记录
  • 高速下载文档
  • 免费文档工具
  • 分享文档赚钱
  • 每日登录抽奖
  • 优质衍生服务

作者详情

相关内容

更多

热门标签

人际关系 动力学连接体 力的合成 配电装置 高考理综 全宋诗作者索引 公务员考试
/ 5
评分 收藏
分享
立即下载
客服
关注