Wilamowski B.M., Irwin J.D. The Industrial Electronics Handbook. Second Edition: Industrial Communication Systems

Подождите немного. Документ загружается.

Automatic Data Mining on Internet by Using PERL Scripting Language 66-9

pages.may.not.be.fully.depictive.of.the.templates.of.all.other.pages..Poor.performance.of.learned.rules.is.

exp

erienced

.for.pag

.tha

.fol

low

.tem

plates

.unc

overed

.by.the.lab

eled

.pa

ges.

. is.pro

blem

.can.be.sol

ved

by.lab

eling

.mor

.pag

es,

.bec

ause

.mor

.pag

.cov

.mor

.tem

plates.

.How

ever,

.man

ual

.lab

eling

.req

uiring

a.la

rge

.su

pply

.of.la

bor

.an

.is.ti

.co

nsuming

.wi

.an.un

satised

.co

verage

.of.al

.po

ssible

.te

mplates.

.met

hod

.of.dat

.ext

raction,

.DMR

.opti

mizes

.the.dat

.min

ing

.pro

cess

.and.mak

.it.bec

ome

.a.

pop

ular

.tool.in.ext

racting

.dat

.fro

.web.pag

es.

.PER

.scr

ipt

.wit

.reg

ular

.exp

ressions

.and.mod

ules

inc

reases

.the.spe

.of.dat

.ext

racting

.as.wel

.as.the.acc

uracy.

. e.DMR.is.cus

tomized

.acc

ording

.to.the.

req

uired

.da

.an

.th

.fo

rmat

.of.da

.th

.us

ers

.de

sire.

references

[CKGS06].C.H..Chang,.M..Kayed,.R..Girgis,.and.K.F..Shaalan,.A.survey.of.web.information.extrac-

tion

. sys

tems,

. IEE

E Transactions on Knowledge and Data Engineering,

. 18(

10),

. 141

1–1428,

Oct

ober

.200

[H99]

.S..Ho

lzner,

.PERL B

lack Book,

.Co

riolisOpen

.Pr

ess,

.Sco

ttsdale,

.AZ,.Edi

tion

.2004.

[KCSG07]

.M..Key

ed,

.C.-H..Cha

ng,

.K..Shaa

lan,

.and.M.R..Girg

is,

.FiVa.tec

.Pag

e-level

.dat

.extract

ion

fro

.temp

late

.pag

es,

.Sev

enth IEEE International Conference on Data Mining Workshops 2007 (ICDM

Workshops 2007),

.Oma

ha,

.NE,.pp

.15–20,.Oc

tober

.28–31,.2007.

[NGW08]

.S..Nee

li,

.K..Gov

indasamy,

.B.M..Wil

amowski,

.and.A..Malin

owski,

.Aut

.dat

.mining.fro

.web.

serv

ers

.using.PERL.scri

pt,

.Int

ernational Conference on Intelligent Engineering System 2008

.(INES

2008),

.Mi

ami,

.FL,.pp.191–196,.Fe

bruary

.25–29,.2008.

[PW09]

.N..Pha

.and.B.M..Wil

amowski,

.IEEE.art

icle

.dat

.extract

ion

.fro

.Int

ernet,

.13th IEEE

Intelligent

Engineering Systems Conference

.(INES 2009

.Ba

rbados,

.Ap

ril

.16–18,.2009.

[PYW10]

.N..Pha

.H..Yu,.and.B.M..Wil

amowski,

.Neural.netw

ork

.train

.thro

ugh

.comp

uter

.netw

orks,

24t

h IEEE International Conference on Advanced Information Networking and Applications 2010,

Per

th,

.Aus

tralia,

.pp..1203–1209,.2010.

[W10]

. B.M..Wil

amowski,

. Design. of. netw

ork

. bas

. so

ware,

. 24th IEEE

International Conference on

Advanced Information Networking and Applications 2010,

.Pe

rth,

.Aus

tralia,

.pp..4–10,.2010.

[WSC05]

.I.-C..Wu,.J.-Y..Su,.and.L.-B..Chen,.A.web.dat

.extract

ion

.descri

ption

.lan

guage

.and.its.imple-

entation,

.29th A

nnual International Conference on Computer Soware and Applications Conference

2005

.(COMPSAC 2005),.Vol..2,.Edinburgh,.U.K.,.pp..293–298,.July.25–28,.2005.

V-1

Outlook

67 Trends and Challenges for Industrial Communication

Systems

. Pet

erPalensky

.................................................................................. 67-1

Introduction. •. Ubiquitous.Global.Connectivity.and.Digital.Identity. •. Vertical.

Int

egration

. •. Hyb

rid

.Lo

cal

.Ne

tworks

.and.Qu

ality

.of.Se

rvice

. •. M2M.

Com

munication

. •. Sca

lability

.in.Ha

rdware

.and.So

ware

. •. Ref

erences

68 Processing Data in Complex Communication Systems. Gerhard Zucker,

Dietmar Bruckner, and Dietmar Dietrich

.......................................................................68-1

Introduction. •. An.Archetype.for.Future.Automation. •. Bottom-Up.versus.Top-Down.

Des

ign:

.Be

havioristic

.Mo

del

.ve

rsus

.Fu

nctional

.Mo

del

. •. Aut

omated

.Me

thods

.fo

.Se

nsor

and.Ac

tuator

.Sy

stems

. •. e.Di

agnostic

.Sy

stem

. •. Int

elligent

.Su

rveillance

.Sy

stems

. •. .

 eHu

man

.Mi

.as.an.Ar

chetype

.fo

.Co

gnitive

.Au

tomation

. •. Ref

erences

67-1

67.1 Introduction

In. the.era. of. ever-increasing. connectivity,.every.possible.industrial.process. is.tapped. with.sensors,.

netw

orks,

.and.autom

ation

.infr

astructure

.to.impr

ove

.eci

ency,

.safe

ty,

.and.tran

sparency.

. is.lead

depe

nding

.on.the.doma

in-specic

.requ

irement,

.to.spec

ic

.tech

nologies

.and.meth

ods,

.just.as.netw

orks

for.mine

.are.die

rent

.from.autom

otive

.netw

orks.

.Some.are.stil

.hopi

.for.“the.one,

”

.univ

ersal

.solu

tion,

.but.that.seem

.a.bit.far-

fetched.

. ere.are,.howe

ver,

.seve

ral

.poin

.wher

.exis

ting

.netw

orks

.will.

expe

rience

.impr

ovements

.towar

.a.more.univ

ersal

.tech

nology.

. is.chapt

.list

.major.chal

lenges

.that.

indu

strial

.com

munication

.sys

tems

.(IC

.wil

.fac

.in.the.nea

.fut

ure:

•

. Ubiq

uitous

.glo

bal

.con

nectivity

.and.dig

ital

.ide

ntity

•

. Vert

ical

.int

egration

•

. Machine-to-machine.(M2M).communication

•

. Scal

ability

.of.har

dware

.and.so

ware

67.2 Ubiquitous Global Connectivity and Digital Identity

Personal.experience.tells.us,.Internet.Protocol.(IP).connectivity.is.ubiquitous..Cell.phones.and.PDAs.

grab.3G.or.Wi-Fi.signa

.and.conn

ect

.to.the.Worl

.Wide.Web.with

out

.much.trou

ble.

. is.delu

sive

simpl

icity

.make

.us.beli

eve

.that.M2M.comm

unication

.enjoy

.the.same.luxu

ry.

. is.is,.howe

ver,

.not.the.

case

.Glob

.conn

ectivity

.in.this.cont

ext

.shal

.mean.mobi

.(but.not.nece

ssarily

.wire

less)

.conn

ectivity

over.a.cert

ain

.regi

.and.mayb

.even.over.bord

ers.

.It.boil

.down.to.plug

ging

.a.devi

.into.some.netw

ork

sock

.or.reg

istering

.it.aga

inst

.som

.wir

eless

.net

work

.wit

hout

.muc

.has

sle.

One

.usua

.prer

equisite

.for.glob

.conn

ectivity

.is.a.subs

cription

.to.a.cert

ain

.type.of.serv

ice

.plan

. e.

serv

ice

.prov

ider

.migh

.be.a.publ

.Inte

rnet

.prov

ider,

.a.tele

phone

.comp

any,

.a.corp

orate

.IT.depa

rtment,

.or.

some.othe

.IT.oper

ator.

. e.subs

criber’s

.iden

tity

.is.typi

cally

.den

.via.a.subs

criber

.iden

tity

.modu

.(SIM

)

Trends and Challenges

for Industrial

Communication Systems

67.1. Introduction.....................................................................................67-1

67.2

. Ubiq

uitous

.Glo

bal

.Con

nectivity

.and.Dig

ital

.Ide

ntity............... 67-1

67.3

. Vert

ical

.Int

egration......................................................................... 67-2

67.4

. Hybr

.Loc

.Net

works

.and.Qua

lity

.of.Ser

vice.......................... 67-2

67.5

. M2M.Com

munication....................................................................67-3

67.6

. Scal

ability

.in.Har

dware

.and.So

ware..........................................67-4

References..................................................................................................... 67-5

Peter Palensky

Austrian Institute

of Tec

hnology

67-2 Industrial Communication Systems

card.or.via.user.name.and.password..Unfortunately,.there.is.no.universal.concept.of.identity.established..

Ever

.tech

nology—and

.even.prov

iders—have

.thei

.own.avor.of.the.two.ment

ioned

.meth

ods.

.indu

strial

.comm

unication

.syst

em,

.hook

.up.to. such.a.prov

ider’s

.infr

astructure

.has.to. deal.

with.a.hete

rogeneous

.mix.of.tech

nologies

.like.IP-P

roxy-Routers,

.dial

-up

.proto

cols,

.digi

tal

.cert

icates,

auth

entication

.proto

cols,

.and.pass

words.

.Unfo

rtunately,

.ther

.is.no.singl

e-sign-on

.(SSO

)

.stan

dard

.for.

ICS.inf

rastructure

.tha

.cou

.be.com

pared

.to.SSO.in.the.IT.wor

ld.

Sun

.Java•.Syst

.Acce

.Mana

ger

.[SUN

05]

.is.an.M2M-

capable

.SSO.solu

tion

.that.prov

ides

.fede

rated

.iden

tity

.(aut

hentication

.and.auth

orization)

.supp

ort

.via.Libe

rty

.Alli

ance

.Phas

.2.and.Secu

rity

.Asse

rtion

.Mark

Lang

uage

.(SAM

L),

.sess

ion

.mana

gement,

.dele

gated

.auth

ority,

.and.on-t

he-y

.audi

.of.auth

entication

.atte

mpts.

Another

. appr

oach,

. also. out. of. the. mobi

. comm

unication

. worl

. come

. from. the. Euro

pean

Tele

communications

.Stand

ards

.Inst

itute

.(ETS

I).

.ETSI.disc

usses

.futu

.M2M.arch

itectures

.in.its.tech

nical

.comm

ittee

.ETSI.TC.M2M#0

.and.call

.SSO.“Ide

ntity

.Fede

rations

.Serv

ice,”

.impl

emented

.by.an.

M2M.ser

vice

.gat

eway

.[ET

SI09].

Traditional

.indu

strial

.netw

orks

.stil

.do.not.have.anyt

hing

.like.this

.but.it.can.be.expec

ted

.that.IT,.tele

com-

munication,

.and.indu

strial

.comm

unication

.will.furt

her

.conv

erge,

.so.that.IT.solu

tions

.will.be.adop

ted

.for.ICS.

67.3 Vertical Integration

ICS.were.and.are.traditionally.used.for.automation,.data.acquisition,.process.monitoring,.and.other.

task

.that.are.“nea

.the.proc

ess.”

.Rece

.tren

.in.prod

uction

.show

.howe

ver,

.an.incr

easing

.link.betw

een

the.shop.oor.and.admi

nistrative

.part

.of.an.ente

rprise.

.Appl

ications

.like.supp

.chai

.mana

gement

and.dig

ital

.eco

systems

.are.rs

.ver

sions

.of.tota

lly

.int

egrated

.busi

ness

.pro

cesses.

.tren

.is.rein

forced

.by.ecol

ogical

.cons

traints

.and.the.incr

eased

.need.for.ener

.eci

ency.

.In.

orde

.to.calc

ulate

.the.carb

.foot

.of.a.prod

uced

.prod

uct,

.it.is.nece

ssary

.to.link.data.from.a.larg

vari

ety

.of.sour

ces.

. is.incl

udes

.auto

mation

.data

.sche

dules,

.supp

lier

.info

rmation,

.and.logi

stics

.data

Even.if.the.vari

ous

.IT.syst

ems

.of.thes

.sour

ces

.(ent

erprise

.reso

urce

.plan

ning

.sow

are,

.manu

facturing

auto

mation,

.logi

stics

.and.eet.mana

gement,

.ener

.cons

umption

.equi

pment,

.etc.

)

.are.some

how

.inte

linked,

.they.usua

lly

.lack.a.comm

.mana

gement

.and.a.comm

.lang

uage

.to.exch

ange

.info

rmation.

Transmitting

.data.“fro

.I/O.to.CEO”.is.ther

efore

.not.only.a.prob

lem

.for.the.tran

sport

.medi

.but.

more.a.ques

tion

.of.sema

ntics

.and.loss

.tran

slations

.betw

een

.IT.syst

ems.

. e.incr

eased

.use.of.IT.tech

nologies

.like.web.serv

ices,

.XML,.and.serv

ice-oriented

.arch

itecture

.(SOA

)

.ease

.this.goal

.Orga

nizations

like.OASI

.[OAS

IS]

.are.enga

ged

.in.appl

ying

.web.serv

ices

.and.XML-b

ased

.form

ats

.to.all.kind

.of.busi-

esses

.and.pro

cesses.

.One.exa

mple

.is.oBI

.[Eh

rl03]

.for.exc

hanging

.bui

lding

.dat

Also,

.the.desig

.and.spec

ication

.phas

.of.indu

strial

.inst

allations

.is.more.and.more.inu

enced

.by.

trad

itional

.IT.meth

ods.

.UML-

.[UML

PA]

.and.OPC.XML.DA.[OPC

DA]

.are.exam

ples

.of.the.succ

essful

usag

.of.IT.met

hods

.for.ind

ustrial

.sys

tems.

True

.inte

rworking

.and.inte

roperability

.cons

ists

.of.a.numb

.of.step

. e.prer

equisite

.is.prot

ocol

.con-

ormance,.i.e.,.using.the.right.plugs,.frames,.addresses,.coding,.timing,.etc..Above.protocol.conformance.

come

.the.appl

ication

.with.its.data.type

.and.func

tional

.pro

les.

.One.impo

rtant,

.alth

ough

.oen.forg

otten

fact

.is.inte

roperable

.netw

ork

.and.appl

ication

.mana

gement

.and.con

guration.

.Whil

.prot

ocol

.conf

or-

mance

.usua

lly

.lead

.to.inte

roperable

.netw

ork

.mana

gement,

.the.appl

ications

.stil

.migh

.use.some.prop

ri-

etary

.way.of.stor

ing

.thei

.con

guration

.para

meters.

.SOA.and.the.use.of.XML-

based

.data.form

ats

.eases.the.

“low

er”

.part

.(i.e.

.data.repr

esentation

.and.enco

ding)

.of.appl

ication

.mana

gement;

.the.seman

tics

.and.usag

of.the.indi

vidual

.para

meters,

.howe

ver,

.must.be.den

.in.stan

dardized

.high

-level

.pro

les.

67.4 Hybrid Local Networks and Quality of Service

Most.ICS.were.designed.with.a.particular.application.domain.in.mind.. ere.are.networks.for.net-

work-based

.cont

rol,

.for.driv

es,

.for.data.acqu

isition,

.mult

imedia

.stre

aming,

.or.re.alar

.syst

ems.

 e.appl

ication

.doma

.resu

lted

.in.spec

ic

.feat

ures

.like.low.late

ncy,

.guar

anteed

.time.slot

.good.

Trends and Challenges for Industrial Communication Systems 67-3

scalability,.or.low.costs..“Universal”.networks.that.can.satisfy.every.possible.connectivity.need.do.not.

exis

.but.hybr

.netw

orks

.are.gett

ing

.clos

.hybr

.netw

ork

.cons

ists

.of.more.than.one.comm

unication

.chan

nel,

.idea

lly

.of.very.dive

rse

chan

nels.

.An.exam

ple

.is.a.mobi

.devi

.with.GPRS.and.GPS..It.combi

nes

.a.bidi

rectional

.best

-eort

chan

nel

.with.a.unid

irectional

.info

rmation

.sour

ce.

.Simi

lar

.thou

ghts

.are.behi

.ener

.mana

gement

node

.tha

.use.Int

ernet

.and.ele

ctricity

.gri

.fre

quency

.as.inf

ormation

.sou

rces

.[FK

07].

Another

.poss

ibility

.to.move.towa

.univ

ersal

.netw

orks

.is.newe

.deve

lopments

.wher

.feat

ures

.of.

seve

ral

.doma

ins

.were.combi

ned.

.Most.nota

bly,

.this.happ

ened

.on.the.medi

.acce

.cont

rol

.laye

.in.the.

case.of.netw

orks

.like.IEEE.1394.or.IEC.6158

.Typi

cally,

.one.part.of.the.band

width

.is.rese

rved

.for.dete

ministic

.tra

.(i.e

.isoc

hronous

.slot

s),

.whil

.the.rest.of.it.can.be.arbi

trated

.in.a.CSMA.way.as.many.

othe

.netw

orks

.do..With.this.hybr

.desig

.two.very.die

rent

.netw

ork

.qual

ities

.can.be.achi

eved:

.Easy.

scal

ability

.for.non

-real-time

.ser

vices

.and.gua

ranteed

.ser

vice

.for.rea

l-time

.app

lications.

.impo

rtant

.facto

.for.hybr

.desig

.is.how.qual

ity

.of.serv

ices

.(QoS

)

.is.impl

emented.

.As.sure

ly,

.the.

IP.suit

.will.play.a.cont

inuously

.impo

rtant

.role.in.ICS,.it.is.good.to.know.that.versi

.6.of.IP.addr

esses

QoS..Alth

ough

.one.(esp

ecially

.an.ICS).can.stil

.feel.v4’s.prob

lems

.in.v6,.it.has.at.leas

.(besi

.its.oen

mentioned

.bett

.scal

ability)

.ow.labe

ling,

.bett

.prio

rities,

.and.is.prep

ared

.for.exte

nded

.head

ers

.that.

migh

.cont

ain

.addi

tional

.QoS.info

rmation.

.Sure

ly,

.the.desig

ners

.of.IPv6.QoS.had.mult

imedia

.serv

ices

in.min

.but.ICS.are.als

.sen

sitive

.to.(al

.fou

.asp

ects

.of).QoS

•

. Band

width

•

. Late

ncy

.(pa

cket

.del

ay)

•

. Late

ncy

.var

iation

.(ji

tter)

•

. Pack

.los

Bandwidth

.migh

.be.the.leas

.impo

rtant

.aspe

.in.80%.of.all.indu

strial

.comm

unication

.(esp

ecially

auto

mation)

.case

.but.robo

.visio

.and.imag

.proc

essing

.is.on.the.rise.at.the.shop.oor

.Also

.the.size.

of.inst

allations

.and.the.numbe

.of.node

.is.dram

atically

.incr

easing,

.so.band

width

.will.play.a.larg

.role.

than.up.to.now.wher

.an.IP.netw

ork

.over.100.Mb.Ether

net

.will.prac

tically

.neve

.expe

rience

.a.band

width

.prob

lem

.in.an.autom

ation

.envi

ronment.

.Late

ncy

.and.jitt

.are.most.impo

rtant

.for.netw

ork-based

cont

rol

.wher

.guar

anteed

.time

.beha

vior

.is.requ

ired.

.Espe

cially,

.time.sync

hronization—a

.very.impo

tant

.serv

ice

.in.dist

ributed

.syst

ems

.(thi

.of.syst

.diag

nostics

.in.proc

ess

.cont

rol

.or.mult

imedia)—

needs

.accu

rately

.den

.uppe

.boun

daries

.for.late

ncy.

.Pack

.loss.is.no.prob

lem

.for.medi

.stre

ams

.but.

coun

ted

.quan

tities

.duri

.prod

uction

.or.emer

gency

.shut

down

.mess

ages

.dema

.high.avai

lability

.and.

reli

ability

.of.the.ICS..It.is.to.be.expe

cted

.that.QoS.will.play.a.more.prom

inent

.and.visib

.role.in.the.

futu

.of.ICS.. e.exis

ting

.prop

osals

.and.impl

ementation

.of.IPv6.QoS.is.cert

ainly

.not.suc

ient

.for.ICS.

and.we.hav

.to.wor

.on.sui

table

.ext

ensions,

.pro

bably

.via.its.ex

ible

.hea

der.

67.5 M2M Communication

Machines.are—and.will.increasingly.be—used.to.act.on.behalf.of.humans..In.the.past,.this.meant.physi-

cal

.labo

.nowa

days,

.it.is.more.and.more.deci

sion

.maki

.and.“so

er”

.task

.Comp

uter

.prog

rams

.that.

act.on.beha

.of.cert

ain

.indi

viduals

.or.role

.are.comm

only

.call

.sow

are

.agen

ts.

.Just.like.huma

ns,

.so-

are

.agen

.need.to.comm

unicate

.with.othe

.agen

ts,

.the.envi

ronment,

.and.huma

.to.be.able.to.ful

thei

.task

.Unli

.huma

ns,

.howe

ver,

.they.are.not.lega

.and.physi

cal

.pers

.with.an.asso

ciated

.iden

tity

and.tru

st.

. is.is.a.ver

.imp

ortant

.dis

tinction

.whe

.it.com

.to.inf

ormation

.sec

urity.

Most

.exis

ting

.secu

rity

.meas

ures

.rely.on.crypto

graphic

.meth

ods

.to.sati

sfy

.secu

rity

.need

.like.auth

en-

ticity

.or.con

dentiality.

. e.secu

rity

.prob

lem

.is.boil

.down.to.prot

ecting

.a.cert

ain

.crypto

graphic

.key.

duri

.setu

.and.oper

ation.

.If.the.end.poin

.of.the.resp

ective

.comm

unication

.rela

tion

.are.huma

ns,

this.key—

which

.prov

ides

.iden

tity

.and.con

dentiality—might

.be.prot

ected

.with.pers

onal

.iden

tica-

tion

.numbe

.(PIN

.or.othe

.“side.chan

nels,”

.whic

.can.only.be.prov

ided

.by.the.inte

nded

.pers

on.

Tran

sferring

.thes

.esta

blished

.proc

edures

.to.M2M.comm

unication

.is.not.triv

ial.

. e.only.aor

dable

67-4 Industrial Communication Systems

method.nowadays.is.using.electronic.chip.cards.(a.k.a..smart.cards),.trusted.platform.modules.(TPM),.

or.othe

.chip

.that.cont

ain

.the.key.and.algo

rithms

.to.use.the.key..With.this

.the.prob

lem

.of.iden

tity

.and.

trus

.is—a

gain—transferred

.furt

her:

.to.the.admi

nistration

.of.manu

facturing

.and.ship

ping

.the.devi

ces.

Each.one.mus

.get.its.own.per

sonal

.chi

.tha

.de

nes

.its.ide

ntity

.and.hel

.wit

.sec

urity.

.seco

.big.chal

lenge

.for.M2M.comm

unication

.is.netw

ork

.mana

gement.

.As.netw

orks

.are.gett

ing

larg

er,

.it.is.requ

ired

.that.trad

itional

.netw

ork

.admi

nistration

.task

.are.autom

ated.

.Assig

ning

.addr

esses

and.baud.rate

.is.alre

ady

.past

;

.most.of.thes

.thin

.happ

.in.a.plug

-and-participate

.way.. e.next.step.

is.plug

-and-work,

.wher

.also.the.netw

ork

.high

.laye

.and.the.appl

ication

.are.con

gured

.autom

ati-

cally.

. e.key.ing

redients

.for.thi

.are

•

. Service.discovery.and.lookup.tables

•

. Self

-description

.of.ser

vices

.and.cap

abilities

•

. Appl

ication

.pro

les

Naturally,

.not.ever

.appl

ication

.can.be.con

gured

.autom

atically

.with

out

.huma

.invo

lvement,

.but.

lear

ning

.net

work

.man

agement

.tool

.can.eas

.tha

.job.dra

matically.

67.6 Scalability in Hardware and Software

 e.trend.for.“total.integration,”.i.e.,.the.as-seamless-as-possible.connection.of.dierent.applica-

tions

.and.netw

orks,

.open

.up.the.reas

.for.the.very.exis

tence

.of.die

rent

.netw

orks.

. eir.die

rent

feat

ures

.that.made.them.suit

able

.for.spec

ic

.appl

ication

.area

.migh

.be.a.barr

ier

.for.inte

gration.

One.of.them.is.size

.Size.in.this.cont

ext

.part

.mean

.the.phys

ical

.size

.but.more.the.code.size.of.pro-

ocol

.stac

ks,

.the.sili

con

.size.for.embe

dded

.func

tionality,

.and.ulti

mately

.the.pric

.tag..Comb

ining

low-

cost

.sens

.netw

orks

.with.powe

rful

.mult

imedia

.or.real

-time

.netw

orks

.does.not.work.out-

of-

the-box.

.What.is.need

.is.a.scal

able

.arch

itecture

.wher

e—in

.term

.of.hard

ware,

.prot

ocols,

.and.so-

are—very

.die

rent

.node

.can.inte

roperate,

.alth

ough

.they.migh

.only.shar

.a.smal

.set.of.comm

serv

ices

.[PAL

04].

One

.rst.step.towa

.such.a.scal

able

.archi

tecture

.is.the.intr

oduction

.of.6LoP

AN,

.a.down

sized

.vers

ion6

IP.over.IEEE.802.

15.4

.(the.popu

lar

.wire

less

.tran

sport,

.also.used.by.ZigB

.and.othe

.tech

nologies).

Othe

.wire

less

.tech

nologies

.like.ISA-

100.11a

.and.Wirel

essHART

.show.simil

.pote

ntial.

.But.the.idea.

shou

.not.be.limi

ted

.to.wire

less

.tech

nologies.

. e.conc

ept

.of.scal

able

.arch

itectures

.must.be.tech

nol-

ogy

.agno

stic,

.ensu

ring

.that.medi

.acce

ss,

.addr

essing,

.mana

gement,

.and.serv

ice

.disc

overy

.foll

.some.

mini

mum

.core.requ

irements

.that.allo

.for.conn

ecting

.a.$0.5.node.to.a.$500.node.in.orde

.to.shar

.some.

func

tionality.

. Inst

rumentation,

. Syst

ems,

. and. Autom

ation

. Soci

ety

. (ISA

)

. den

. six. clas

ses

. of. indu

strial

comm

unication:

•

. Clas

.0:.Eme

rgency

•

. Clas

.1:.Clo

sed-loop

.reg

ulatory

.con

trol

•

. Clas

.2:.Clo

sed-loop

.sup

ervisory

.con

trol

•

. Clas

.3:.Ope

n-loop

.con

trol

.(hu

man

.in.the.loo

•

. Clas

.4:.Ale

rting

•

. Clas

.5:.Log

ging,

.dow

nloading

Class

.0.fall

.into.the.cate

gory

.safe

ty,

.the.othe

.clas

ses

.are.eith

.cont

rol

.(1.and.2),.moni

toring

.(4.and.5),.

or.some

thing

.in.betw

een.

.Depe

nding

.on.its.clas

.a.cert

ain

.serv

ice

.or.appl

ication

.assu

mes

.cert

ain

QoS,.and.oper

ating

.appl

ications

.with.die

rent

.clas

ses

.over.one.and.the.same.tran

sport

.is.not.easy

.ISA-

00.11a

.sugg

ests

.using.IEEE.802.

15.4

.as.physi

cal

.and.TCP/

UDP/IPv6

.as.netw

ork

.and.tran

sport

.laye

rs.

For.Clas

.0,.howe

ver,

.even.wire

.netw

orks

.are.oen.not.suc

ient.

.It.migh

.stil

.be.a.long.way.unti

.a.real.

univ

ersal

.and.sca

lable

.net

work

.is.ava

ilable.

Trends and Challenges for Industrial Communication Systems 67-5

references

[Ehrl03].P..Ehrlich.and.K..Sinclair,.What.is.oBIX?,.AutomatedBuildings.com,.December.2003.

[ETSI09]

. Sop

hia

. Ant

ipolis,

. Init

ial

. pro

posed

. arc

hitecture

. for. M2M. netw

orks,

. wor

king

. docum

ent

M2M01_026.of.ET

.TC.M2M#01,.Ja

nuary

.2009.

[FK07]

.F..Kup

zog

.and.P..Palen

sky,

.Wide-ar

.cont

rol

.system

.for.bal

ance-energy

.pro

vision

.by.energ

cons

umers,

.Pr

oceedings of IFAC FET 2007,

.To

ulouse,

.Fra

nce,

.2007.

[OA

SIS]

.Organiza

tion

.for.the.Adva

ncement

.of.Str

uctured

.Info

rmation

.Sta

ndards,

.Avai

lable

.at.www

.oasis-

open.org,

.Acces

sed

.on.Augu

.2010.

[OPCD

.OPC.Foun

dation,

.OPC.XML-DA.1.00.spe

cication,

.Versio

.1.01,.OPC.Foun

dation,

.Jul

.2003.

[PAL04]

. P.. Palen

sky,

. Req

uirements

. for. the. next. genera

tion

. of. buildin

. netw

orks,

. Pro

ceedings of

International Conference on Cybernetics and Information Technologies,

. Syst

ems and Applications

(ISAS CIT

SA 2004),

.Orl

ando,

.FL,.2004.

[SUN05]

.M..B..Bai

kie

.and.S..Gaede,. e.role.of.mob

ile

.opera

tors

.and.the.JAV

A•

.pla

tform

.in.the.machin

to-machine,

.Whi

.Pa

per,

.Sun.Micr

osystems,

.San

.Cl

ara,

.CA,.Jun

.2005.

[UMLP

.Spe

cication

.of.har

.rea

l-time

.indu

strial

.aut

omation

.system

.wit

.UML-P

.Whit

.Pap

er,

Univ

ersity

.of.St

uttgart,

.St

uttgart,

.Ger

many,

.Ma

.2005.

68-1

68.1 Introduction

In.automation,.mechanical.units.are.more.and.more.replaced.by.electronic.components.. ey.con-

tain

.sens

ors,

.actu

ators,

.and.info

rmation

.comp

uting

.syst

ems.

. e.goal.is.to.achi

eve

.high

.proc

ess

.and.

prod

uct

.qual

ity.

. is.mean

.e.g.

.to.incr

ease

.the.perf

ormance

.by.impl

ementing

.more.func

tionality.

Syst

ems

.beco

.inte

rconnected,

.whic

.caus

.incr

eased

.comp

lexity—something

.that.cann

.easi

be.hand

led

.with.trad

itional

.tech

nological

.appr

oaches.

.It.is.nece

ssary

.to.reth

ink

.exis

ting

.appr

oaches

and.nd.new.solut

ions.

. ese.can.be.foun

.in.stat

istical

.mode

ls,

.hier

archical

.mode

ls,

.and.bion

.mod-

ls.

. is.chap

ter

.revi

ews

.rece

.rese

arch

.resu

lts

.about.meth

ods

.to.redu

.the.eor

.for.comm

ission,

main

tenance,

.and.oper

ation

.of.the.syst

em.

.Stat

istical

.mode

.are.used.to.autom

atically

.deri

.nor-

.oper

ation

.cond

itions.

.Hier

archical

.mode

.help.in.abst

racting

.info

rmation

.from.the.mere.sens

valu

.to.sema

ntic

.conc

epts

.that.are.also.used.in.huma

.thin

king

.and.conv

ersation.

.In.the.bion

rese

arch

.dire

ction,

.cont

emporary

.rese

archers

.like.Mark.Solm

.and.Anto

nio

.R..Dama

sio

.buil

.upon.

psyc

hoanalysis,

.whic

.coul

.be.such.a.new.appr

oach

.for.engi

neers.

. e.theo

ries

.are.widel

.unkn

own

to.engi

neers,

.stil

.thei

.meth

ods,

.whic

.orig

inate

.from.neur

ology,

.are.part

.simi

lar

.to.the.meth

ods

.of.

comp

uter

.tech

nology.

Processing Data in

Complex Communication

Systems

68.1. Introduction.....................................................................................68-1

68.2

. An.Arc

hetype

.for.Fut

ure

.Auto

mation..........................................68-2

68.3

. Botto

m-Up

.ver

sus

.Top

-Down

.Desi

gn:

.Beh

avioristic

Mode

lversus

.Fun

ctional

.Mod

el....................................................68-2

68.4

. Autom

ated

.Met

hods

.for.Sen

sor

.and.Act

uator

.Sys

tems.............68-4

68.5

. e.Dia

gnostic

.Sys

tem....................................................................68-4

Error.Detection. •. Statistical.Generative.Models. •. Online.

Parameter.Updates. •. Results

68.6. Intelligent.Surveillance.Systems....................................................68-7

Architecture

68.7. e.Human.Mind.as.an.Archetype.for.Cognitive.

Autom

ation.......................................................................................68-9

Perception.in.Automation:.A.Historic.Overview

References...................................................................................................68-10

Gerhard Zucker

Vienna University

of Tec

hnology

Dietmar Bruckner

Vienna University

of Tec

hnology

Dietmar Dietrich

Vienna University

of Tec

hnology

68-2 Industrial Communication Systems

68.2 an archetype for Future automation

If.we.want.to.be.able.to.handle.complex.control.systems.that.we.are.facing.today.in.automation—and.

whic

.has.been.the.driv

ing

.forc

.behi

.arti

cial

.inte

lligence

.for.the.last.50.year

s—then

.we.shou

.lear

to.unde

rstand

.how.the.huma

.mind.work

.since.it.is.the.best.cont

rol

.syst

.that.we.have.avai

lable.

Espe

cially

.the.high

er,

.more.abst

ract

.leve

.are.rele

vant;

.proc

essing

.in.the.lowe

.laye

rs—the

.neur

ons—are

in.this.cont

ext

.less.inte

resting,

.since.a.tech

nical

.solu

tion

.will.not.grou

.on.a.biol

ogical

.base

.it.will.

inst

ead

.use.ele

ctronic

.com

puters.

Using

.a.prop

.mode

.of.the.huma

.mind.as.an.arch

etype

.allo

.us.to.desig

.tech

nical

.syst

ems

.that.

are.able.to.cont

rol

.comp

lex

.proc

esses:

.surv

eillance

.syst

ems

.for.airp

ort

.secu

rity

.to.dete

.dang

erous

scen

arios;

.eci

ently

.inte

rpreting

.vast.amou

nts

.of.data.from.thou

sands

.of.embed

ded

.syst

ems

.in.a.buil

ing

.or.in.a.prod

uction

.plan

.prov

iding

.supp

ort

.in.a.hous

ehold

.and.espe

cially

.supp

orting

.the.elde

rly

.in.

geri

atric

.appl

ications;

.cont

rolling

.trai

.and.airp

lanes.

. e.prin

ciple

.behi

.all.thes

.appl

ications

.is.

iden

tical:

.a.mach

ine

.has.to.perc

eive

.its.envi

ronment

.and.inte

rpret

.the.perc

eived

.info

rmation

.on.a.high

abst

ract

.lev

.and.dra

.rea

sonable

.con

clusions.

.mode

.that.can.desc

ribe

.huma

.beha

vior

.is.not.suc

ient.

.A.huma

.bein

.is.much.too.comp

lex

.that.we.

coul

.infer.its.func

tion

.dire

ctly

.from.its.beha

vior.

.It.will.neve

.reac

.a.stat

.that.is.iden

tical

.to.a.stat

.in.the.

past

—there

.are.no.two.iden

tical

.stat

.in.one.life

. e.ques

tion

.for.an.appr

opriate

.func

tional

.mode

.is.also.not.

answ

ered

.in.the.desc

ription

.of.syna

ptic

.conne

ctions

.betw

een

.neur

.stru

ctures,

.sinc

.this.desc

ription

.does.

not.expl

ain

.how.a.huma

.“fun

ctions,”

.i.e.

.it.does.not.expl

ain

.the.func

tional

.modu

les

.that.it.cont

ains.

.To.give.

an.anal

ogy:

.if.we.have.a.desc

ription

.of.the.beha

vior

.of.a.word.proc

essing

.prog

ram,

.it.will.not.suc

.to.exam

ine

.the.transistor.structures.of.the.CPU.on.which.it.runs;.we.will.not.be.able.to.describe.the.functionality.of.

a.word.proc

essing

.prog

ram,

.too.much.comp

lexity

.lies.betw

een

.the.oper

ation

.of.a.sili

con

.chip.and.a.sow

are

appl

ication.

.Stil

.an.elec

tronic

.comp

uter

.syst

.is.by.far.simp

ler

.than.a.biol

ogical

.brai

.we.look.back.at.the.deve

lopment

.of.arti

cial

.inte

lligence

.(AI)

.we.see.the.grea

.achi

evements

.that.

have.been.prod

uced.

.Norb

ert

.Wiene

.was.alre

ady

.awar

.that.natu

.is.the.sour

.of.answ

ers.

.We.have.to.

be.rea

.to.acc

ept

.res

ults

.fro

.oth

.sci

entic

.dis

ciplines.

But

.we.also.real

ize

.that.we.have.made.some.big.mist

akes

.in.AI.and,.in.our.opin

ion,

.some

times

.le.

the.grou

nds

.of.scie

nce.

. is.shou

.now.be.corr

ected

.by.allo

wing

.alte

rnative

.consi

derations

.besid

.the.

trad

itional

.app

roaches.

68.3 Bottom-Up versus top-Down Design:Behavioristic

Model versus Functional Model

According.to.[ENF07].and.[Pal08].we.can.distinguish.four.generations.of.AI:.First,.symbolic.AI,.which.

work

.on.symbo

lic

.codi

ng,

.mani

pulation,

.and.deco

ding

.of.info

rmation.

.Symbo

lic

.proc

essing

.meth

ods

are.toda

.an.ess

ential

.bas

.for.AI,.whi

.is.als

.sho

.in.the.wor

.of.[Ve

l08].

Statistical

.AI.empl

oys

.netw

orked

.stru

ctures

.and.thei

.coup

lings.

. e.cent

ral

.topic.is.lear

ning,

.whic

unfo

rtunately

.alwa

.only.lead

.to.one.or.more.optim

.but.does.not.go.furt

her.

.Lear

ning

.from.what.has.

been.lear

ned

.cann

.be.mode

led—the

.syst

.is.not.able.to.chan

.its.lear

ning

.stru

ctures

.base

.on.what.

has.been.lear

ned.

. is.is.also.unde

rstandable

.from.a.die

rent

.view

point,

.since.only.the.abil

ity

.to.per-

anently

.store.imag

.and.scen

arios

.and.inte

grate

.them.into.the.lear

ning

.proc

ess

.can.yiel

.the.desir

perf

ormance

.and.div

ersity.

.We.wil

.dis

cuss

.tha

.a.bit.fur

ther

.bel

ow.

Next,

.we.real

ize

.that.the.body.with.its.needs.and.its.requ

ests

.for.sati

sfaction

.is.a.neces

sary

.requ

irement

for.inte

lligence—this

.is.the.disci

pline

.of.embo

died

.inte

lligence.

.Inte

lligence

.does.not.end.in.itsel

.it.serve

the.body.to.subs

ist

.eci

ently,

.as.Anto

nio

.R..Dama

sio

.stat

.in.[Dam9

4].

.Seen.from.the.phil

osophical

view

point

.this.was.grea

.prog

ress,

.becau

.this.step.had.sign

icant

.conse

quences

.on.the.discu

ssions

.of.the.

exis

tence

.of.a.soul.(Heg

els

.body

-mind

.prob

lem).

.If.we.work.scien

tically,

.it.becom

.clea

.that.thin

king

and.body.are.the.same.thin

.only.seen.from.die

rent

.view

points.

. is.is.exac

tly

.the.stat

ement

.of.mode

psyc

hoanalysis

.and.wher

.neur

ology

.meet

.psyc

hoanalysis

.and.merg

.into.neur

opsychoanalysis

.[Sol

02].

Processing Data in Complex Communication Systems 68-3

Finally,.at.the.end.of.the.1980s,.it.was.understood.that.the.brain.has.to.be.more.than.just.a.system.on.

whi

.di

erent

.pro

grams

.run.for.di

erent

.pur

poses

.[Pa

l08].

.Emo

tions,

.lan

guage

.and.con

sciousness

.are.

pro

perties

.tha

.are.of.gre

.rel

evance

.for.hum

.bei

ngs.

. is.is.the.dom

ain

.of.emo

tional

.AI..Fir

st,

.the

was.no.sepa

ration

.bet

ween

.emo

tions

.and.fee

lings

.and.we.sim

ply

.ass

umed

.tha

.the

.are.eval

uation

.mech

nisms.

.Sti

ll,

.phe

nomena

.lik

.con

sciousness

.or.the.sens

ation

.of.fee

ling something

.cou

.not.be.exp

lained.

Analysis

.of.thi

.his

torical

.dev

elopment

.mak

.two.thi

ngs

.cle

ar.

.Fir

st,

.we.see.tha

.eng

ineers

.hav

.cho

sen

.a.bot

tom-up

.met

hod

.for.the

.desi

gns,

.simp

.due.to.lac

.of.kno

wledge.

.We.put.one.pie

.on.top.of.

the.oth

.to.wor

.our.way.fro

.neu

rons

.(wh

ich

.wou

.be.equ

ivalent

.to.har

dware

.in.com

puter

.mod

els)

to.con

sciousness

.(wh

ich

.is.the.hig

hest,

.mos

.abs

tract

.lev

el).

.But.we.sho

uld

.not.do.tha

.we.sho

uld

.alw

ays

wor

.top-

down

.lik

.com

puter

.eng

ineers

.are.use

.to.wor

.Sec

ond,

.we.hav

.sel

ected

.cer

tain

.str

uctures,

fun

ctions

.or.beh

avior

.fro

.the.who

.com

plex

.of.hum

.abi

lities,

.and.tri

.to.imp

lement

.the

.in.tec

nical

.sys

tems.

.We.exp

ected

.to.und

erstand

.the.res

.whi

.we.le.out

.in.goo

.tim

.Esp

ecially,

.the.las

two.gen

erations

.of.AI.hav

.use

.psy

chological

.kno

wledge,

.whi

.was

.how

ever,

.not.che

cked

.for.con

sistency.

. is.res

ulted

.in.a.pat

chwork—a

.mix.of.di

erent,

.par

tly

.con

tradicting

.sch

ools

.tha

.wer

.all.

mer

ged

.toge

ther

.into.one.mod

el.

.Suc

.an.exa

mple

.is.sho

.in.[Br

e02],

.whi

.vio

lates

.sci

entic

.pri

ciples.

.If.res

ults

.of.di

erent

.psy

chological

.sch

ools

.are.use

.whi

.hav

.not.bee

.che

cked

.for.con

sistency

(or.in

teroperability,

.to.us

.a.mo

.te

chnical

.te

rm),

.th

.re

sults

.ar

.wo

rthless.

.mus

.hav

.imp

lications

.for.fut

ure

.res

earch,

.whi

ch,

.in.our.opi

nion,

.lea

.to.the.h.gen

eration

of.AI.. e.bou

ndary

.con

ditions

.are.cle

ar:

. 1..Com

puter

.en

gineers

.mu

.de

mand

.top

-down

.de

sign.

. 2.. We.ne

.a.un

iform

.mo

del,

.wh

ich

.is.fr

.of.co

ntradictions.

. 3..Eng

ineers

.ha

.to.co

operate

.wi

.sc

ientists,

.wh

.ha

.al

ways

.de

alt

.wi

.th

.br

ain

.an

.th

.mi

nd.

Computer

.eng

ineers

.suc

cessfully

.use.the.lay

ering

.sho

.in.Fig

ure

.68.

.Suc

.a.lay

.mod

.is.cur

rently

.not.ava

ilable

.for.the.min

.but.we.hav

.to.try.to.app

.the.lay

.mod

.any

way.

.If.we.wor

.top-

own,

.how

ever,

.it.bec

omes

.cle

.tha

.an.abs

traction

.of.the.neu

ral

.net

work

.[CR

04]

.(th

.har

dware,

.to.

com

pare

.it.wit

.the.com

puter

.dom

ain)

.is.not.nec

essary

.at.the.beg

inning;

.we.wil

.exp

lain

.tha

.a.bit.

fur

ther

.bel

ow.

.We.do.not.kno

.how.the.bra

.is.“pr

ogrammed”

.on.the.neu

ral

.lev

el.

.And.eve

.if.we.

kne

.it.wou

.not.be.the.rig

.lev

.of.mod

eling

.(ju

.thi

.of.ali

gning

.tra

nsistors

.unt

.you.get.a.wor

pro

cessing

.app

lication);

.we.hav

.to.foc

.on.the.psy

che,

.and.we.hav

.to.rem

ember

.tha

.mod

eling

.is.a.

mul

ti-layer

.pro

cess

.(Fi

gure

.68.

1).

. er

efore

.we.can

not

.map.spe

cic

.beh

avior

.of.a.pro

cess

.(an.“ap

plica-

tion”

.in.the.com

puter

.dom

ain)

.dir

ectly

.to.fun

ctional

.mod

ules.

.In.doi

.so.we.wou

.red

uce

.the.com

plex

sys

tem

.“hu

man

.bei

ng”

.to.a.pri

mitive

.sys

tem,

.whi

.con

tradicts

.rea

lity:

.the.con

trol

.uni

.of.a.hum

bei

.doe

.not.exe

cute

.the.sam

.beh

avior

.twi

.abs

olutely

.ide

ntical,

.sinc

.con

ditions

.alw

ays

.cha

nge

[Ja

a97].

.Tod

ay’s

.rob

ots

.(wh

ich,

.in.thi

.con

text,

.are.the.sam

.as.pup

pets)

.hav

.one.thi

.in.com

mon:

.the

are.not.hum

an,

.but.it.is.the.hum

.obs

erver,

.who.pro

jects

.hum

an-like

.pro

perties

.into.the

.Acc

ording

to.[So

l02],

.hum

.bei

ngs

.con

sist

.phy

siologically

.and.psy

chologically

.of.man

.di

erent

.con

trol

.loo

FIGURE 68.1 Layer.model.of.the.brain.with.reference.to.a.computer.