System-level feature-based modeling of cyber-physical systems

(1)

Delft University of Technology

System-level feature-based modeling of cyber-physical systems

A theoretical framework and methodological fundamentals

Pourtalebi, Shahab DOI 10.4233/uuid:a75df963-aede-44a7-9216-274cdc7c4278 Publication date 2017 Document Version Final published version Citation (APA)

Pourtalebi, S. (2017). System-level feature-based modeling of cyber-physical systems: A theoretical framework and methodological fundamentals. https://doi.org/10.4233/uuid:a75df963-aede-44a7-9216-274cdc7c4278

Important note

To cite this publication, please use the final published version (if applicable). Please check the document version above.

Copyright

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons. Takedown policy

Please contact us and provide details if you believe this document breaches copyrights. We will remove access to the work immediately and investigate your claim.

This work is downloaded from Delft University of Technology.

(2)

(3)

(4)

System-level feature-based modeling of

cyber-physical systems

A theoretical framework and methodological fundamentals

Proefschrift

ter verkrijging van de graad van doctor

aan de Technische Universiteit Delft,

op gezag van de Rector Magnificus Prof.ir. K.C.A.M. Luyben;

voorzitter van het College voor Promoties,

in het openbaar te verdedigen op Woensdag 26 April 2017 om

10:00 uur

Door

Shahab POURTALEBI

Master of Science in Industrial Design

(5)

This diss

promoto

Compos

Rector M Prof. dr.

Indepen

Prof. dr. Prof. dr. Prof. dr. Prof. dr. Dr. S. Bo Dr. F. De Prof. dr.

System

A theor

Keywor

features;

ISBN/EA

An elect

http://r

Copyrig

shahab

sertation h

or: Prof. d

sition of

Magnificus, I. Horváth

ndent me

J-P Pernot I.S. Sariyil F.J.A.M. v G.W. Kort orgo erakhshan A.R. Balke

m-level f

retical fra

rds:

Cybe ; system-le

AN:

978-tronic ve

repository

ght © 201

b60p@gm

has been

dr. I. Horv

the docto

h,

embers:

t ldiz an Houten uem enende

feature-mework a

r-physical evel model

94-028-0

rsion of t

y.tudelft.n

17 by S. P

mail.com

approved

áth

oral comm

chairp Techn Arts e Techn n Unive Techn Cons Unive Techn

based m

and meth

system; sy ing; pre-em

0621-2

this disse

nl/

Pourtaleb

d by the

mittee:

person nische Univ et Métiers nische Univ ersity of Tw nische Univ iglio Nazio ersity of Ta nische Univ

modeling

hodologic

ystem man mbodimen

rtation is

bi

versiteit D ParisTech versiteit D wente versiteit D nale delle abriz versiteit D

g of cybe

cal fundam

nifestation f t design; m

available

elft, promo elft elft Ricerche elft

er-physi

mentals

features; s mass custo

e at

otor

cal syste

system-lev omization;

ems

el

(6)

There are The tradit ing and t tions [1]. operation ment, wh cessing. entails th resembla and cybe and appr cyber-phy practicall As a succ premise C cal and so processes nantly by decision systems w energy, t tion and the overa modeling support i intrinsic c control, a distillatio among ot CPC-base the run-t tion and a the physi of the ne and foreig and Inno competit European focus on

1.1

e several in tional cons heir place i Typically th n. Their imp hich includ In the field hat many of ance in term erware tech roach of di ysical com y all softwa cessor of th CPC offers i ocial enviro s and havin y systems making cap will interac asks, and p automation all comput g approach in order to challenges adaptive re on, robustn thers [4]. ed systems ime obtain adaptation ical world t ew opportu gn research ovation of tiveness of n Institute cyber-phys

PROLIFER

teresting tr sumer prod s taken ove he new pro plementatio des novel f d of techno f the new p ms of their hnologies is igital comp puting (CP are integrat he concepts is that digit onment. Th ng the con which are pabilities [3 t with hum performanc n in decisio ting system es and too o be able t originate i easoning, a ess, system gather dat ned data in , and provi hrough sop nities offer h initiatives the Europe enterprise of Innovat sical system

INT

1 RATION OF

rends in th ducts (so-ca er by new k oducts are d on is facilita functional m ologies, we roducts are r functions. s a strong t puting eme C). This als ted and con s of ubiqui tal comput is integratio ntrol algorit equipped 3]. The gov mans, physic ce dimensio on making, m. CPC also ls, design p to take all in technolo and timed m troublesh ta from the quasi-real de results t phisticates red by CPC s, such as th ean Comm s and SME ion and Te ms, which r

TRODUCT

F CYBER-PH

e field of p lled consum kinds of pro digitally co ated by the materials, e e can obse e exploiting . The conve rend in our erged a de so has a str ntrolled con tous and p ing will be on was des thms of co with the n verning visi cal systems ons. Thoug humans w o entails th principles, d the challen ogical issue actuations hooting, en e physical w time, deve that control hardware a have been he Horizon 2 mission, the

Es, and the echnology, epresent a

TION

HYSICAL C

product dev mer durabl oducts and/ nnected an e current ra energy sou erve a proc g the same c ergence of r days [2]. I ecade ago, rong effect nsumer dur pervasive co much mor scribed also mputation necessary n on is cyber s, and the e gh CPC pav

ill still play he need fo developme nges succe es such as , but in pr nergy prov world throu elop strateg l objects, su and softwar n set as an o 2020 Frame e COSME 2 e Strategic just to na new parad

COMPUTIN

velopment es) are grad /or artifact-nd offer som apid techno urces, and cess of con core techno f digital ha n addition, which has t on the im rables. omputing, e integrate o as penetra defined in networking r-physical c environmen ves the way

an importa r complete ent paradig essfully. Ho distributed rivacy, secu isioning, an ugh various gic plans fo ubjects and re actuator objective in ework Prog 2014-2020 Innovation me a few. digm in pro

G

and distrib dually disap -service com me level of ological adv informatio nvergence, ologies and rdware, so , a novel co s been nam mplementat the fundam ed with the ation into re n run-time, g, reasoning computing-nt in time, y to decent ant role as p ely new the ms, and run owever, the d sensing, s urity, inform nd sustaina s sensors, p or optimal o d other syste s [5]. Explo n many Eur gram for Re program fo n Agenda f These pro oduct and s bution. ppear- mbina-f smart vance-n pro-which d show ftware oncept med as tion of mental physi-eal life domi-g, and -based space, traliza-part of eories, n-time e most system mation ability, process opera-ems in itation ropean search for the for the ograms service

(14)

develo reinte ubiqu duces and im curren and e wirele matio driven the bo As ind comp intrins identi funda nical s for fo happe er, if w and d tial co windo etc. cy numb aspec tal for are at From cally hetero ble o adapt (vii) ba relatio the p ment for ge From tion, w adapt al com ing, e opment. T egrates diffe uitous comp s brand new mplementa nt technolo electronics, ess networ on processin n by cyber-oundaries o

CY

1.2

dicated abo ponents, bu sically diffe ifies synerg amental cha systems. Th orms, which en if we loo we enforce escribe tho onstituents, ow, medica yber-physic ber of these cts the syste r any system least to be an architec distributed ogeneous c r adaptive ted internal alance in te ons, and (vi

rinciples of of large sc eneration ne an operati while also p ting/self-ev mponents, xecution a The underp erent conce puting” [8] w concepts ation. Essen ogical prog and the ev ked senso ng, and sem -physical co of the curre

YBER-PHYS

ove, CPSs a t systems t rent constit gy as the m aracteristics his is an imp h often hid ok at these e an abstrac ose characte , rather tha al, military, cal systems e abstract c em is looke ms, in orde e considered ctural aspe d and dece component system bo l architectu erms of the ii) multi-lev f architectu ale CPSs, u ew compon onal aspec providing m olving capa (iv) harmon nd control, pinning pa epts and p [9], and “th s of system ntially, the p ress and p volution of rs and act mantic kno omputing, c ent system d

SICAL SYST

are not just hat intends tuents in an main charact s of CPSs t portant issu de their gen systems at ct view of t eristics that an its specif transportat s may show characterist d at. Since r to identify d by model ect, CPSs ar entralized ts, (iii) high oundary, (v ure, (vi) high amount of vel compos ure driven se behavio nents [15]. t, CPSs are means and abilities, (iii nized opera , computin radigm of aradigms s he Internet m architecti paradigm o hysical min network sy tuators, Int wledge eng conceptual design and

TEMS AND

a combina s to achieve n interconn teristics of that differe ue since CP nuine distin t the level o the differen t concern t fic compon tion, aerosp w similar sy tics can be architectur y the main l-based CPS e branded architectur h level of sy v) hetarach h level of he f hardware, ition of sca developme or as constra characteriz utilities, (ii) ) intense in ating princ g and reas cyber-phy such as “em of things” ing and no of CPS is see niaturizatio ystems, em ternet of (e gineering i lization of C architectin

THEIR MA

ation of sep e a rather ti nected conc CPSs [14]. T ntiate them PSs systems nguishing of their tech nt manifest he whole o nents and e pace, agric ystem level e very large ral and ope characteris S developm by the foll re, (ii) ad-h ystem comp hically orga eterogeneit software a ales. These c ent approa aint for inte

zed by: (i) s ) external f nformation iples with r soning, and ysical syste mbedded sy [10]. Never ovel forms en by many n in the fie mbedded so every)thing n the cybe CPSs requir ng arena.

AIN CHARA

parate phys ight couplin cept [12] [1 There are h m from oth s also manif characteris hnical impl tations of C of systems a elements. T ultural, pro l characteri e dependin erational asp stics of CPS ment approa owing cha hoc functio plexity, (iv) anized or r ty with reg nd cyberwa characterist ches, whic erfacing, an services ori unctional a exchange a regards to d activating em (CPS) e ystem” [6] [ rtheless, it a of system y as the res elds of mec oftware and gs, large-sc r domain [ res steppin

ACTERISTIC

sical and co ng among t 3]. Some re however m her enginee fest in a wid stics. This m ementation CPSs, we ca and/or thei Through the oduction, ho istics. Obvio g on the n pects are fu Ss, these tw aches. racteristics: onal conne fully open, runtime dy ards to com are compon tics are add

h facilitate nd indicate ented over adaptability among the sensing an g and motiv essentially [7], “smart also intro-operation sult of the chatronics d systems, cale infor-11]. Being g beyond

CS

omputing these two esearchers many more ered tech-de variety may easily n. Howev-n ideHowev-ntify r substan-e abstract ome care, ously, the number of undamen-wo aspects : (i) physi-ections of , extenda-ynamically mponents, nents and dressed by develop-e thdevelop-e ndevelop-edevelop-ed rall opera-y and function-nd observ-vating, (v)

(15)

growing smart co proactivit context s approach opment o (of system The impa CPS clarif which can ing from and persp make the mation p compone obtained tions in a

A N

1.3

The typic new desig principles instance, cyber-phy tion and nents, wh mization, cyber-phy cyber com ing a sof laws as w the relate portance unit of tim various lo statistical describin In the co proposed approxim ed devel Raghav, G ture desc to associ Heteroge functional ontrol acco ty and auto sensitive in hes aim at of individua ms). They ty artial consid fy several o n be an issu the adven pectives of em disting processing c ents, (iv) d , and reaso n unsuperv

NEW CHAL

cal high lev gn challeng s of design as Grimm, ysical syste representa hich is usua , and verifi ysical syste mponents) ftware arch well as the i ed sensors of modelin me that is c ocal param l propertie g the intera ontext of ad d to describ mations of c lopment a G., & Gopa cription, to iate it with eneity and intelligenc rding to g onomous o teraction w implement al systems ypically use deration of opportuniti ue for resea t of CPSs th system con uishable a capability, ( ifferent so oning based vised mann

LLENGE: CR

vel of hete ges in the f and mode , C. and Ou ems at syste tion of who ally accomp ication[18]. ems to all c must be ac itecture wh mplementa or actuato ng workloa communica eters but a es of the w actions amo dvancing m be the com computatio nd integra lswamy, S. use constra h behavior complexity e extendin goal-driven operation, (v with human ting these c and theref architectu integration es that cou arch. The in hat has nev nsideration

re: (i) hybr (iii) fixed or urces of p d sources, a ner [17].

ROSS-BOU

erogeneity, field of syst eling CPSs h u, J. pointed em level an ole of the s panied by t Woo, H. e classes of f chieved by hich embe ation and v ors [19]. Bo ad (i.e. the ated betwe lso macros workload a

ong the com model-base mputation i onal semant ation of ha proposed aints to sele ral models, y of this kin g to contex operationa viii) capabi n in the loo characterist fore inadeq re as constr n of digital uld be utiliz nteresting p ver been ex s. The most rid structur r ad-hoc fun roviding k and (v) abi

NDARY AR

complexity em enginee have been d d at it, a pa nd as a who ystem inclu the challen et alias arg failures, (em a top-down ds the imp verification gdan, P. an amount of een various copic metr and adopti mponents ed design o n an abstra tics [21]. To ardware, so to prefer n ect combin data, req nd of system xt awarene al scenario lity for uns op. Behavio tics, while a quate for an raint for int

cyber and a zed for diffe point is the xperienced t important re of comp nctional co nowledge lity of learn

RCHITECTIN

y, and inte ering. Chal discussed i articular cha ole is the fo uding both nges of des gued that r merging in n approach plementatio of the soft nd Marcule f measured CPS nodes rics) and de ng a statis [20]. of CPSs, Mo act manner o facilitate t oftware an non-functio nations of th uirements ms have be

ess and sma os, (vii) abi

supervised or driven C address fun nalyzing lar terfacing [1 analogue p erent purp totally new d before in t characteri ponents, (ii nnection (o including ning from h

NG & OPER

egration of lenges, issu n different allenge rais ormal and a physical a ign space e resilience o hardware, h with respe ons of the f tware for po escu, R. em and/or pro s and which esign optim stical phys osterman, P r and to co he concurr nd cyberwa onal analysi he compon and other een identifi art reasonin lity of a le learning, a CPS develop nctionality rge scale sy 6]. physical par oses, and t w situation different a istics of CPS ) real-time or both) be built-in, se history and

RATION DE

CPSs intro ues, method publication sed by des abstract spe nd cyber co exploration of mission-c software a ect to the d feedback c ossible failu phasized th ocessed da h affects no mization bas ics approa P.J. and Zan onsider mo

ent but dis are compo is using arc nent variant r propertie ied as one ng, (vi) evel of nd (ix) pment devel-ystems rts in a threats result-spects Ss that infor-etween ensors-d

situa-ESIGN

oduces ds and ns. For igning ecifica- ompo-n, opti-critical and/or design-control ures in he im-ata per ot only sed on ch for nde, J. deling tribut-onents, chitec-ts, and s [22]. of the

(16)

most challenging engineering issues in many scientific publications [23] Heterogeneity and complexity issues are especially critical in the pre-embodiment design phase, where system designers have to deal with both system compositionality and component com-posability questions. Usually the comcom-posability issues are explicitly addressed in concep-tualization and architecting of CPSs, and subsequently, homogenous constituent design and development are delegated to the experts of the related fields. For example, devel-opment of the software constituents is delegated to software engineers, control engineers deal with control issues, mechanical engineers solve the hardware problems, and electron-ic experts design and develop circuits. Since they are working on a manageable and homogeneous part of CPS, they can tackle the design challenges by using their expertise. However, the compositionality issues of CPSs remain unsolved and challenging for system designers.

A large number of efforts have been made in research and system development to provide methods and techniques for addressing the concrete technical challenges. For instance, Rajhans, A. et alias suggested using parameters in architectural views to support hetero-geneous design and verification [24]. Fallah, Y.P., and Sengupta, R. proposed to design systems in a top-down manner, by identifying system metrics such as approximate measures of awareness, and then characterizing component models and their interactions in the process of designing vehicle safety networks [25]. Zhang, J. et alias discussed the importance and application of modular verification of dynamically adaptive systems [26] Pinto, A., and Krishnamurthy, S. considered uncertainty as key characteristic of cyber-physical system and proposed analysis and design tools for specification and detailing these type of systems in an industrial setting [27] Zilic, Z., and Karajica, B. discussed various high-level design and synthesis challenges and proposed an approach with the emphasis on the numerical values passed between sensors and the rest of the system [28].

It must be noted that the above cited works are just examples of the wide ranging efforts, which have been made to address the numerous technical challenges of cross-boundary architecting and operational design of CPSs. The problem becomes more complicated when consideration of some kind of design intentions and/or optimization for multiple concerns are added to the problem. Design for X (e.g. customizability, maintainability, sustainability, interaction) traditionally has some sort of design principles and/or guide-lines which are not performing well for CPSs due to the unique characteristics of CPSs. Therefore, designing CPSs in the pre-embodiment phase needs new principles, tools, techniques, or a kind of design supporting tool. The traditional ‘reductionist’ handling of the design and implementation concerns is not advantageous from multi-aspect optimiza-tion of complicated and complex systems.

A traditional way of addressing the problem of handling complexity and heterogeneity in pre-embodiment design phase is to apply abstraction and to simplify model of the system in a level which could be handled logically. Due to the abstraction, the complexity of the CPSs is reduced, and heterogeneous components can be represented as simple elements. In fact, the system designers are being pushed to formulate the abstract relationship among components. We refer to this formulation as logic-based modeling. The logic of the systems’ functions could be modelled and simulated through the defined methods of operation (usually as mathematic equations) with available modeling and simulation tools (e.g. SysML, Simulink, Mathematica, LabVIEW). Many software tools and modeling lan-guages have been developed recently for these purposes. These tools and lanlan-guages are

(17)

able to c simulate ty. Multip developm Another to compo ability ch each field during sy sufficient All availa ware, sof commun others, w the scient • Curren to the not no • The a physic tions. afterw • Interdi tional • Compo compo ramete In the lin design sy structure automate automati (actually, modeling technolog intelligen these and challenge Therefore informati not only f ical desig capture req it at a high ple optimiz ment of the method to ose a team allenges of d could exp ystem desi t knowledg ble mentio ftware, and ication. In which can c tific knowle nt modeling necessary s oticed durin bstraction cal attribute Ignoring p wards. isciplinary c point of vie osability co osability ch ers, constra ne of the c ystem, whic d interactio ed (a.k.a. in on of the would ha g systems t gy of mac nce-based c d applying e that cann e, the targe ionally inte functional a gn. uired logic level of ab zation algo system. handle the of designe f architectin plain to the gn. Usually e about all

DEF

1.4

oned soluti d cyberwa addition, t ause misun edge gaps a g tools and simplificati ng system a used for es, dimens physical con considerati ew. oncerns rem hallenges i aints, and va urrently w ch facilitate on of the s telligent) d design pro ave pulled to an artifi chine learn creative pro them in SM not be add et system m erconnected and logical c and relati straction to orithms ha e heteroge rs with vari ng and con e other tea y, the lead related kno

INITION O

ons suffer are compon the termino nderstandin are listed as d technique on imposed architecting simplificat ions, const nsideration ions remain main unsol n CPSs arc alues as we idespread e system-lev system desi design syste ocess wou the resear cial intellig ing-as-a-se oblem solv MFs-based dressed me manifests i d modules design of t ons among o avoid dea ve also be neity of CP ious fields o nceptualizat am membe er of such owledge an

F THE RESE

from the t nents nega ology of ea ng and mis s: es do not s d by the av g and conce ion purpos traints, card s in system n unsolved ved in mos chitecting r ell as the op CAD/CAE s vel feature-igner and t em. Behind ld have op rch topic f gence drive ervice (MLa ing method system de eaningfully tself in a t of function the system, g elements aling with c een develo PS in pre-em of expertise tion of CPS ers which is a team sh nd engineer

EARCH PRO

ruth that in atively affe ach discipli sinterpretat ufficiently s ailable tool eptualizatio ses elimina dinality) in m engineeri both from st system m requires co perational o systems, we -based mod the suppor this decisio pened a ne from the fi en system aaS), and d dologies ar esign would y in one si ool box of nal compon , but also its

s of a syste omplexity a oped to as mbodimen e in order t s. In this wa ssues shou hould be t ring fields.

OBLEM

ntrinsic diff ect proper ine differs tion. The p support CP ls, all techn on. ates physic favor of lo ing cause s m technical modeling t onsidering ones. e considere deling on a rting system on was the ew dimens ield of eng developme deep learni re available d have mea ngle PhD f multiple, p nents. The m s structural em to mod and hetero sist design t design ph o tackle co ay, the exp ld be cons he one wh ferences of interdiscip from that roblems re PS designer ical problem cal matter ogical cons serious pro and comm tools. Solvin architectur ed an inter a goal drive m, rather th e recognitio sion for res gineering s ent). Thoug ing and ar e, combinat ant an eno research p procedural modules su l and morph el and genei-ners in hase is mpos-erts of idered ho has f hard-plinary of the esult in rs. Due ms are s (e.g. sidera-oblems munica-ng the ral pa-ractive en and han an on that search system gh the rtificial tion of rmous project. ly and upport

(18)

holog-We ha easies menti buildi beside risks a the ou Accor impos mode suppo and d retica mode menta ally lis ries to comp by com The o design embo RQ1. RQ2. RQ3. RQ4. RQ5. RQ6. RQ7. RQ8. RQ9. RQ10. RQ11. ad two alte st way wou ioned prob ng a novel es other co and consid utcomes wo rdingly, we sing purely eling framew

ort it. These detailed exp l works res elling frame ation of the sted as und o fill the kno putational c mparing th

1

bjective of ners of het odiment des What is th of CPSs in Why do th tackle com How to co design an What kind geneous What is th Which pie to compu How the p building b How the p computat What is th . What is th blocks? . What are ernative wa ld be to im blems. The m l theoretica onsideratio ering all di ould open a targeted a y physical v work (SFM e theories s ploring the sulted in in ework. The e introduce derstanding owledge ga onstructs f he results w

THE O

1.5

this resear terogeneou sign phase. he state of t n the pre-em he convent mposability onsider all r nd developm d of inform CPSs? he informat eces of info utationally m pieces of in blocks? pieces of in tionally? he computa he procedu the aspect ys to choo mprove the c more diffic al framewo ns such as ifficulties, w a new door a theoretica view in mo F). For dev hould be m knowledge ntroducing computatio ed modellin g the know ap, introduc for impleme ith other av

OBJECTIVE

ch has bee us-complex .” According the art and mbodiment tional desig y issues? required in ment in pre ation shou tion structu ormation sh model com nformation nformation ational proc re of creati s of compa se in order current mo ult way wa rk for CPS complexit we have ch r in system al framewo odeling CP eloping SF made based e gap, and computat onal constr ng framewo wledge gap cing the co entation of vailable mo

ES AND KEY

n formulate x-systems i

gly, the key the knowle t design ph gn methods terdisciplin e-embodim ld to be con ure of the m hould be ca mponents an are related and the rel cedure of c ng CPS mo rison of CP to address odeling tool as to study modeling t ty and hete hosen the s modeling a ork and me Ss, named MF, we nee d on the kn the investi ional const ructs provid ork. The ste p, exploring oncept of m f the framew odeling too

Y RESEARC

ed so as: “to n order to y research q edge gaps i hase? s not suffici nary and mu ment design nsidered fo modeling un ptured by t nd constitu together in ations amo reating mo odel by com S modeling s the menti ls in order t the main c that consid erogeneity. econd way arena. ethodologic as system eded unde owledge p igation of c tructs that de sufficien eps towards g causalities modelling fra work, and f ls in the fra

CH QUESTI

o develop a o model th questions a in designin ently suppo ulticultural n phase? or uniformly nits? the knowled ents of CPS n order to c ong them ca odeling buil mposing mo g tools and ioned prob to adapt th challenges a ders physica Being awa y with the h cal fundam m-level featu erpinning th rovided by causalities. are neede nt support f s this end a s, construct amework, p finally benc mework lev

ONS

a supportin hese system re listed as: g and deve ort CPS des aspects of C y modelling dge frames Ss? create mod an be captu lding block odeling buil languages? blems. The em to the and to try al matters are of the hope that mentals for ure-based heories to y technical The theo-ed for our for imple-are gener-ting theo-proposing chmarking vel. ng tool for ms in pre-: elopment signers to CPS g hetero-s in order eling ured ks? lding ?

(19)

RQ12. What are the advantages and disadvantages of the introduced modeling frame-work in pre-embodiment design of CPSs, in comparison with the current modeling frameworks?

Our research plan was designed to answer the mentioned key questions in five research cycles (RCs). For this reason, the key questions are grouped and assigned to research cycles. The first research cycle aimed to answer first two questions. The questions three and four are assigned to the second research cycle. Focus of the research cycle three was on the questions five to seven. The aim of research cycle four was to answer questions eight to ten. And finally, the questions eleven and twelve are answered in the fifth research cycle. Table 1-1 summarizes the relations of research questions to the chapters.

The objective of the first research cycle was to explore the state of the art about designing CPSs in the pre-embodiment design phase. The concerned studies were done in the following contexts: (i) overall characteristics of CPSs, (ii) relevant design approaches, (iii) deriving the required design principles, (iv) relevant design-assistive tools, and (v) ad-vantages of feature technology.

The objective of the second research cycle was to develop a theoretical framework for interdisciplinary and multicultural system-level modeling of CPSs in the pre-embodiment design phase. This theoretical framework should support uniform modeling of heteroge-neous CPSs through specifying the kinds of information which should be captured. The theoretical work was supported by empirical studies.

The objective of the third research cycle was to provide computation methodological fundamentals and information constructs to underpin modeling of CPSs. This also investi-gated the concept and implemented system-level manifestation features. The information structures, which specify the required pieces of information for computational modeling of CPSs, as well as the relations among the pieces of information, were the main contexts of the studies performed in this research cycle.

In the fourth research cycle, the objective was twofold: (i) testing computational feasibility of the results of the theoretical works, and (ii) moving towards implementation of the introduced modeling entities and procedures through designing information schema constructs and computational processes for the proposed modeling tool. Computationally

Table 1-1 Mapping of the research key questions to the research cycles

Chapter 2 Chapter 3 Chapter 4 Chapter 5 Chapter 6

RQ1 Research cycle 1 RQ2 RQ3 Research cycle 2 RQ4 RQ5 Research cycle 3 RQ6 RQ7 RQ8 Research cycle 4 RQ9 RQ10 RQ11 Research cycle 5 RQ12

(20)

captu inform them, tions CPSs, of this The o cycle conce again guage study priate of CP ties o the c mode cordin aspec This r flow o above two ty resear (DIR), an exp for co protot inquir In the outco develo activit and d detail throu The se explo constr quirem develo ring the mation and , and the of a tran were the m s research c objective o was to ben eptual m st the ava es in the fo in order to eness for pr Ss. The stu of the prop characterist eling tools ng to the cts and indic esearch is of activitie ementioned ypes of fram rch in desig and a ben ploration p onverting types). The ry process. e first resea ome of the oped reaso ties are per document a about the gh focus gr econd rese rative part ructing the ments resu opment. T required d the relatio procedural nsparent m main work cycle. of the fift nchmark th modeling ailable tool orm of a c o measure t e-embodim udy conside posed solut tics of th and lang derived c cators. part of a b s shown in d research ming which gn context chmarking art and a co the availa e implemen arch cycle, forerunnin oning mod rformed th analyses. Th e knowledg roup discus earch cycle started w e conceptu lted in ide hese theor chunks o ons among considera modeling o ing contex th research e proposed framework ls and lan comparative their appro ment design ered novel tion agains he curren guages, ac comparison

RE

1.6

background n Figure 1-questions, h have been t (RIDC), fo study cond onfirmative ble knowle ntation of th a multi-asp ng literature del, five do rough focu he theory re ge gap. Fo ssions with concentra ith empiric ual framew ntifying the ries underp of g -of xt h d k -e o-n -t t c-n

ESEARCH A

d research p -1. This pro has been d n introduce llowed by ducted bas e part. The D edge into he construc pect reason e study an mains of s used literat esulted from or some do experts. ted on dev cal studies work. Study e classic th pinned dev Figure

1-APPROACH

project, wh omotion re decompose ed in [29] [3 three cycle sed on RIDC DIR frame a concepts ctive part s ning mode d knowled study have ure study, m the explo omains of s velopment in order to ying availab heories that velopment -1 Process research

H

hich procee esearch, aim ed into five 30]. These i es of desig C. All resea also include (and/or th serves as re l was deve ge synthes been iden as well as orative stud study, the of concep o identifyin ble theorie t could sup of the con flow of the b h project eds accordi ming to an research cy nclude one gn inclusive arch cycles es a constru he develop esearch mea eloped base sis. Accordi ntified. The web-based dy provide theory is c tual framew ng requirem s related t pport our fr nceptual fr background ng to the nswer the ycles with e cycles of e research consist of uctive part pment of ans in the ed on the ng to the e research d searches sufficient confirmed work. The ments for to the re-ramework ramework d

(21)

named as the cases The third with dem manifesta theory is ture revie part of th principles The fourt part. The are form construct addition tive part based on The fifth against th most pro performin on the re by quest analyzed In the fo each chap chapters research s mereo-op s provided i d research c monstration ation featu indeed an ew on syst he research s support th th research concept of ulated as t ting databa procedures of this rese n critical rea research c he available oper model ng mechan equirement tionnaire an and conclu llowing pa pter, and th as researc cycle. perandi the n the foreru cycle comb n in the cho ures (SMFs) extension em modeli focuses on hem. h cycle also f informatio theoretical ases (DB) of s of SMF cr earch cycle soning and cycle aimed e modeling ing tools a nism. Moreo s of pre-em nswered b uded as the ragraphs, t he relations ch cycles. I ory (MOT). unning em bined theor osen cases. was made and comp ing theorie n developin o consists o on schema developm f genotype reation and e. The resul d consistenc d to bench g tools and and languag over, in para mbodiment y the expe e confirmati

TH

1.7

the thesis o s between t t also brie The theore pirical stud retical deve In fact, the e through lementary es helps to ng knowled of theoretic construct ent part. T (GT) and p d model co ts of const cy analysis. marking of languages. ges as well allel the asp modeling erts involve ion of the w

HESIS OUT

outline is re the chapter efly represe etical devel dies. elopment o e confirmat case studi of MOT. Ra develop th ge frames f al develop and proces The introdu phenotype mposition tructive par . f the devel . The study l as collecti pects of co of CPSs. Th ed in mod whole resea

TLINE

epresented rs. Figure 1-ents condu opment ha of the infor tion of the es and ana ational reas he theory. for modelin ment part ss of step-w uced theori (PT) as wel are conside rt are used loped mod started wit ing informa mparison a he compari eling CPSs arch work. d by explain -2 visually d uct of the as been test rmation str theory of s alyses. The soning and The constr ng unites a and constr wise SMF cr ies are test

l as model ered as con for confirm deling fram th identifyi ation abou are derived son is perfo s. The resu ning object demonstrat research in ted on ucture system e SMFs litera-ructive nd the ructive reation ted by DB. In nstruc-mation mework ng the ut their based ormed lts are tive of tes the n each

(22)

The m eral p Five d featur gap fr not ad tools f Chapt It star system been know mode funda ponen Theor tion a aspec struct know Chapt tation MOT mode amon of SM house The a theore pared accord Comp ison a naire Chapt and su To Ch Pourta the co main object phenomeno domains of re technolo rom differe dequate fo for support ter 3 aimed rted by emp m compone explored. ledge gap. eling comp amentals ne nts of CPSs. ry of system and extensi cts of CPSs a ture provid ledge fram ter 5 explai nally. The a and SMFs els have be ng them as MF creation es. aim of Chap etical effort d with our p ding to the paring the m aspects on answered b ter 7 concl uggestions hapter 2: alebi, S., Ho ontext of c tive of Chap on and to p f CPS, desi ogy are stud ent perspec or designin ting pre-em d to propos pirical stud ents, availa According This theory lex system eeded for u . m manifesta ion of MOT and their co ded by SM es. ns how the im was to theories. T een design proposed b and mode pter 6 is to ts through proposed to e characteri modeling fr the basis o by modelin udes the re for follow

1

orváth, I., & cyber-phys

pter 2 was provide a c gn approa died for pro ctives. It is g CPSs. Th mbodiment se a concep ies. After id ble theorie gly, mereo y underpins s to system uniformly f ation featur T, SMFs the omponents MF theory e informatio propose co Three data ned that re by the SMFs el compos o present t benchmar ool on fram stics of CPS rameworks of the resul g experts. esearch by up research

RELAT

.8

& Opiyo, E. ( sical consu to acquire comprehen ch, design oviding com proved tha e reasons design of C ptual framew dentifying r s that could -operandi s the rest o m-level feat formulating res is introd eory aims to s and const is repres on structur omputation abases of g eflects the s theory. Th ition in co the outcom rking. Five mework leve Ss and the have been ts of analyz y mentionin hes.

TED OWN

(2013). “Mu mer durab fundamen sive and ri principles, mprehensiv at the conv of insuffici CPSs are als work for pr requiremen d support o theory is of our resea ture-based g interdisci duced in th o offer form tituents in a ented in re of SMFs c nal construc genotypes, pieces of his chapter nnection w me of the v modelling el. The aspe requireme n completed zing a set o ng the mai

PUBLICATI

ulti-aspect s bles”, In AS tal insights gorous des , design su ve insight a ventional d ently of th so explored re-embodim nts of mode our concept introduced rch by shift modeling. plinary and he Chapter mal descrip a uniform w architectur could be im cts to supp phenotype informatio also introd with the en validation o tools were ects of comp nts of pre-e d according of indicato n findings,

IONS

study of m SME 2013 I s in the stu scriptive kn upporting t bout the kn design prin e current m . ment desig eling hetero tual framew d for fillin ting the pa . MOT intro d multicult 4. As comp ption for al way. The inf

ral and op

mplemente port feasibi es, and ins n and the duces the pr ntry form a of the resu e chosen to parison we embodimen g to specific rs and the research o ass custom Internation died gen-nowledge. tools, and nowledge ciples are modelling n of CPSs. ogeneous work have ng in the radigm of oduces all ural com- plementa-l required formation perational d compu-lity of the stantiated relations rocedures and ware-ults of the o be com-re chosen nt design. c compar- question-outcomes, mization in nal Design

(23)

Engineering Technical Conferences and Computers and Information in Engineering Con-ference, Portland, Oregon, USA. pp. V004T05A006-V004T05A006.

Pourtalebi, S., Horváth, I., & Opiyo, E. Z. (2014). “New features imply new principles? Deriv-ing design principles for mass customization of cyber-physical consumer durables”, In Proceedings of the 10th international symposium on tools and methods of competitive engineering TMCE 2014, Budapest, Hungary. pp. 95-108.

To Chapter 3:

Pourtalebi, S., Horváth, I., & Opiyo, E. Z. (2014). “First steps towards a mereo-operandi

theory for a system feature-based architecting of cyber-physical systems”, In Proceedings

of GI-Jahrestagung, ACDP 2014, Stuttgart, Germany. pp. 2001-2006.

Horváth, I., & Pourtalebi, S. (2015). “Fundamentals of a mereo-operandi theory to support

transdisciplinary modeling and co-design of cyber-physical systems”, In Proceedings of the

ASME 2015 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, Boston, Massachusetts, USA. pp. V01AT02A005-V01AT02A005.

To Chapter 4:

Pourtalebi, S., and Horváth, I., (2016), "Towards a methodology of system manifestation features-based pre-embodiment design", Journal of Engineering Design, Vol. 27 (4-6), pp. 232-268.

To Chapter 5:

Pourtalebi, S., and Horváth, I., (2016), "Information schema constructs for defining warehouse databases of genotypes and phenotypes of system manifestation features",

Frontiers of Information Technology & Electronic Engineering, Vol. 17 (9), pp. 862-884. Pourtalebi, S. and Horváth, I. (2016), “Procedure of creating system manifestation features:

An information processing perspective”, In Proceedings of the 10th international

symposi-um on tools and methods of competitive engineering TMCE 2016, Aix-en-Provence, France. pp. 129-142.

Pourtalebi, S. and Horváth, I. (2016), “Information schema constructs for instantiation and

composition of system manifestation features, Frontiers of Information Technology &

Electronic Engineering, Vol. 17 (x), pp. xxx-xxx.

To Chapter 6:

Pourtalebi, S., & Horváth, I. (2016d). Benchmarking the conceptual framework of a system-level manifestation features-based toolbox, (in preparation for a journal publication), pp. x-x.

(24)

[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] Horváth, I. D., (2011), and servic Gantz, J., a shadows, a pp. 1-16. Horváth, I. Describe, E Springer, p Khaitan, S. cyberphys Horvath, I. physical sy Design, Lin Mechatron Edwards, S embedded 85 (3), pp. Wolf, W.H. the IEEE, V Prekop, P., Computer Want, R., F worlds wit Proceedin Atzori, L., I Networks, Sha, L., Go frontier", P Anonymo Foundatio Lee, E.A., (2 FL, pp. 363 Lee, E.A., a Systems A Doddapan (2012), "A sensor net Software E Zheng, X., developm Horváth, I. and imple Behrendt, Grimm, C., systems", P Woo, H., Y developm Beijing, pp Bogdan, P design opt ., Rusák, Z., M "An informa e combinati and Reinsel, and biggest ., (2014), "Wh Explain and pp. 99-120. .K., and McC sical systems ., (2012), "Be ystems", Pro nz (Austria), nics. S., Lavagno, d systems: fo 366-389. ., (1994), "Ha Vol. 82 (7), pp , and Burnet r Communic Fishkin, K.P., th electronic gs, pp. 370-3 era, A., and Vol. 54 (15), opalakrishna Proceedings us, (2006), "C on, 2006. 2008), "Cybe 3-369. and Seshia, S Approach, 1s neni, K., Ever model-drive tworks", Proc Engineering Julien, C., Po ent for cps a ., and Gerrits mentation p M. (Eds.), Or , and Ou, J., ( Proceedings Yi, J., Browne, ent method p. 525-528. ., and Marcu timization",

1.9

Martin, E.H., V ation techno ions", Procee D., (2012), "T growth in th hat the Desig Predict?", in Calley, J.D., (2 s: A survey", yond advan ceedings of 27-29 June, L., Lee, E.A., ormal mode ardware-soft p. 967-989. tt, M., (2003) ations, Vol. 2 Gujar, A., an c tags", Conf 377. Morabito, G , pp. 2787-28 n, S., Liu, X., of the, Taich Cyber-Physic er physical sy S.A., (2011), I t ed., LeeSes r, E., Gemikon en engineeri ceedings of for Sensor N odorozhny, application", sen, B.H.M., ( principles", P rganizing Co (2011), "Unif s of the, San , J.C., Mok, A dology for re ulescu, R., (20 IEEE Design

REFERE

9

Van Der Veg ological spec edings of the The digital u he far east", gn Theory o : An Antholo 2015), "Desig IEEE System ced mechat the Proceed 2012, ACCM and Sangiov ls, validation tware co-des , "Activities, 26 (11), pp. 1 d Harrison, B ference on H ., (2010), "Th 805. and Wang, Q hung, pp. 1-9 cal Systems" ystems: Desi ntroduction shia.org. nakli, O., Ma ing framewo the Proceed Network App R., and Casse , Technical R (2012), "Cybe Proceedings ommittee of fying proces Diego, CA. A.K., Atkins, E silient cyber 011), "Cyber and Test of

ENCES

gte, W.F., Koo cification of a e, Vol. 9, Wa niverse in 20 IDC iView: ID f Social-Cyb ogy of Theor gn technique ms Journal, Vo tronics: new dings of the M Austrian Ce vanni-Vincen n, and synthe sign of embe context and 1168-1176. B.L., (1999), " Human Facto he Internet o Q., (2008), "C 9. ", (NSF), offic ign challeng n to Embedd lavolta, I., M ork for archit dings of the T plications, IE ez, F., (2014) Report UTAR er-Physical S of the TMCE TMCE 2012, ss networks f E., and Xie, F. r-physical sy physical syst Computers, oijman, A., O abstract pro ashington, D 020: Big data DC Analyze t er-Physical S ries and Mod es and appli ol. 9 (2), pp. 3 design chal 1st Worksho enter of Com ntelli, A., (19 esis", Procee edded system d ubiquitous "Bridging ph ors in Compu of Things: A s Cyber-physic cial web site ges", Proceed ed Systems: ostarda, L., a tecting and a Third Interna EEE Press, pp ), "Braceasse RISE-2015-00 Systems: con E, Horváth, I. , Germany, p for design of ., (2008), "De ystems", Proc tems: Worklo , Vol. 28 (4), p Opiyo, E.Z., an ototyping for DC, pp. 209-2 a, bigger dig the future, V Systems Mus dels of Desig cations of 350-365. lenges of so op on Mecha mpetence in 997), "Design edings of the ms", Proceed s computing hysical and v uting System survey", Com cal systems: A of National S dings of the, : A Cyber-Ph and Muccini analysing wi ational Work p. 1-7. rtion: Behav 02, 2014. ncepts, techn ., Rusák, Z., A pp. 19-36. f cyber phys esign and ceedings of t oad modelin pp. 78-87. nd Peck, r artifact 223. gital Vol. 2007, st gn, cial-cyber-atronic n of e IEEE, Vol. dings of ", virtual ms - mputer A new Science , Orlando, ysical , H., ireless kshop on vior-driven nologies Albers, A., sical the, ng and

(25)

[21] Mosterman, P.J., and Zander, J., (2011), "AdvancingModel-Based Design by Modeling Approximations of Computational Semantics", Proceedings of the Proceedings of the 4th International Workshop on Equation-Based Object-Oriented Modeling Languages and Tools; Zurich; Switzerland; September 5; 2011, Linköping University Electronic Press, pp. 3-7.

[22] Raghav, G., and Gopalswamy, S., (2010), "Architecture driven development for cyber physical systems", SAE International Journal of Aerospace, Vol. 3 (1), pp. 95-100.

[23] Fallah, Y.P., Huang, C., Sengupta, R., and Krishnan, H., (2010), "Design of cooperative vehicle safety systems based on tight coupling of communication, computing and physical vehicle dynamics", Proceedings of the, Stockholm, pp. 159-167.

[24] Rajhans, A., Bhave, A., Loos, S., Krogh, B.H., Platzer, A., and Garlan, D., (2011), "Using parameters in architectural views to support heterogeneous design and verification", Proceedings of the, Orlando, FL, pp. 2705-2710.

[25] Fallah, Y.P., and Sengupta, R., (2012), "A cyber-physical systems approach to the design of vehicle safety networks", Proceedings of the, Macau, pp. 324-329.

[26] Zhang, J., Goldsby, H.J., and Cheng, B.H.C., (2009), "Modular verification of dynamically adaptive systems", Proceedings of the, Charlottesville, VA, pp. 161-172.

[27] Pinto, A., and Krishnamurthy, S., (2010), "Developing design tools for uncertain systems in an industrial setting", Proceedings of the, Monticello, IL, pp. 1714-1721.

[28] Zilic, Z., and Karajica, B., (2011), "High-level design of integrated microsystems - Arithmetic perspective", Proceedings of the, Montreal, QC, pp. 77-82.

[29] Horváth, I., (2007), "Comparison of three methodological approaches of design research", Proceedings of the Proceedings of International Conference on Engineering Design ICED, Vol. 7, pp. 28-31.

[30] Horváth, I., (2006), "On the differences between research in design context and design inclusive research", Proceedings of the The International Design Research Symposium.

(26)

2 C

This ch focuse the m knowl atic in issue, aspect system discipl variou studie source Driven rences tives o survey aspect art of which the stu domai The in suppo compu of syst Finally the im

Chapter 2 OVER

hapter prese ed on descri main factors edge gap, w vestigation implying the ts and persp m design and lines. Theref s domains d phenome es. n by the abo s of the phe of the survey y. The reaso ts to be con cyber-physi were consid udy is desig in of the sur nvestigations orting tools a utational de tem-level fea y, sub-chapt mplications o

2 RVIEW

ents the wo bing the stu s influencin which was al and multi-a e necessity o pectives. A v d engineerin fore, the ma of research enon and ble

ove objectiv nomenon. S y as well as t ning model sidered. The ical systems, dered in the gn approach

vey was des

s of these t and the find sign and mo eatures, whic er 2.9 summ f the finding

OF TH

orkflow and udied pheno g this phe lready conci aspect elabo of investigat very broad d ng have bee in objective that have ending the k ve, a survey Subchapter 2 he reasonin identified f e first doma s, as the cate e research pr hes, which ar sign principl three doma ings are pre odeling tools ch was consi marizes and gs in the con

E STAT

results of t omenon and enomenon a isely circums oration. Fillin ting and eva domain of r en put into t e of the first contributed knowledge is performe 2.1 provides g model wh five knowled in, elaborate egory of syst roject. Discu re applied in

les and guid

ins were co esented and s directed ou dered as the blends the ntext of the r

TE OF T

he first rese d the associa and their c scribed in th ng in the kn aluating the esearch inte the center o research cy d to the de and insights ed to explore s further info hich facilitate dge domain ed upon in S tems, the de ssed in Subc n design and delines, whic omplemente analyzed in ur attention e fifth doma findings of research.

THE AR

earch cycle o ated knowle correlations he Introduct owledge ga recent adva erests had to of research in cle was defi scription an s obtained f e and descr ormation ab ed the syste ns for the st Subchapter esign and c chapter 2.3, d developm ch are discus ed by the in Sub-chapte to the emer ain of our stu

the investig

RT

of the resea dge gap, as and causa tion, needed ap was a mu ances from a o be consid n multiple in ined as iden nd explanat from these k

ribe the vari out the spec matic execu udy and spe

2.2, is the s ustomizatio the second ment of CPSs ssed in Subc nvestigation er 2.5. The st rging resear udy in Subc gations, and rch, which well as on alities. The d a system-ulti-faceted all relevant dered since nterrelated ntifying the ion of the knowledge ous occur-cific objec-ution of the ecified the tate of the n issues of domain of s. The third hapter 2.4. of design tudy of the ch domain hapter 2.6. d discusses

(27)

2 OVERVIEW OF THE STATE OF THE ART

2.1 INTRODUCTION

Aim of the first research cycle was to (i) define the concerned knowledge domains of interest, and specifically to (ii) specify the addressed research phenomenon, (iii) describe the observed scientific/engineering knowledge gap related to it, and (iv) landmark the preferred direction of the promotion research. Initially, this research cycle started with the

study of the wide topic of designing of cyber-physical systems, but it has gradually been

narrowed down to the investigation of system-level feature-based modeling of

cyber-physical systems as a result of the refinement applied by means of the insights obtained in the first research cycle. This was an implication of studying a very large number of recent academic research documents and industrial developments in the domains related to our research topic. The main guiding research question has been stated as:

What conceptual frameworks, computational approaches, and implementation methodol-ogies are needed for the theoretical and practical realization of system-level feature-based modeling of cyber-physical systems?

Designing cyber-physical systems as a technological, epistemological and methodological concept has multi-faceted aspects. These aspects could be explored from multiple per-spectives such as design methods, principles and tools, as well as CPS characteristics and other issues associated with design and modeling of CPSs. The studied phenomenon has

been formulated as advanced computer support for designing of cyber-physical systems in

the pre-embodiment phase of development. The characteristics of CPSs have a strong effect on all domains of knowledge and on the aspects of investigation concerning this phenomenon. It was a challenge to find the proper focus, arguable scope, and approach of the research. After getting more information about the unknown, the conduct of the research has been reinterpreted, narrowed down and streamlined multiple times. This thesis however includes only the part of the work that are in the mainstream of our inquiry and conceptualization, and those results which are seen as valuable outcome of this work. Since the investigated phenomenon is multi-disciplinary (i.e. requires the knowledge and methods of multiple disciplines) it was unavoidable to find an exact focal point, which enabled us to make decision on the relevant research aspects, and to clarify the domains of interest for knowledge aggregations. The backbone of the knowledge aggregation was a structured literature survey for exploring the state of the art in the regarding domains. The knowledge domains considered for the study are explained by the reasoning model

presented in sub-chapter 2.1.2. In a nutshell, the studied domains were (i) cyber-physical

systems, (ii) system design approaches, (iii) design principles, (iv) design supporting tools,

and (v) feature-based design. The survey of these domains resulted in a rather

comprehen-sive overview of the state of the art, to define a specific knowledge gap that has been addressed in our research. The findings of the survey also helped us to define the specific objectives for the research with the intent of creating scientific and engineering novelties. In simple words, the knowledge gap has been formulated as lack of knowledge (genuine

(28)

concepts) and proper supporting tools for CPS designers in the phase of pre-embodiment. The knowledge gap was addressed from the abovementioned five perspectives.

2.1.1 Objectives of the first research cycle

The first research cycle aimed at constructing and formulating a descriptive theory that identifies the boundaries of the knowledge gap, determines what part of it can be ad-dressed in this research, and fills in the identified part of it with new knowledge. The orientation studies have made it obvious that there is a relatively large knowledge gap

related to advanced computer support for designing of cyber-physical systems in the

pre-embodiment phase of development. Consequently, objective of the first research cycle was to study (i) the five identified knowledge domains, and synthesize a body of knowledge that describes what the current situation is, and (ii) the break out points re-garding the development of novel tools. There were many useful source materials found in academic publications, but also in professional repositories on the web. However, in some fields the research was found to be still in an immature state. This especially true for the second generation (smart) CPSs. The design principles and supporting tools were more exposed and advanced in the case of the first generation of CPSs, which have their roots in, e.g. embedded systems, smart robotics, digital networking, systems control, and so forth. In addition to the term cyber-physical systems, many similar phrases or nominations have been used in the literature to describe systems with similar or resembling paradigmatic characteristics: ‘advanced embedded systems’, ‘smart robotics’, ‘intelligent sensor net-works’, ‘complex adaptive systems’, ‘industrial Internet of things’, and ‘smart connected everything’. This resulted in an unclearness, which could not be resolved by the different definitions mushrooming in the last years. As a result, many seemingly relevant publica-tion, claiming focus on and contribution to CPSs, were ignored. On the other hand, it made necessary for us to specify own definitions that correctly and exactly conveying the inter-pretations used in this thesis. Interestingly, the overwhelming majority of publications presented mono-disciplinary approaches, although the domain of CPSs is claimed to be multi-disciplinary in nature. It is fair to mention that the term cyber-physical system was used in many publications just because of its novelty and the imperative trends.

2.1.2 Reasoning model

Based on the identification of the pertinent knowledge domains, a simple reasoning model

was formed. This is shown graphically in Figure 2-1. This model indicates a kind of flow of

knowledge through the interrelated domains. This flow conveys and blends the knowledge associated with ‘cyber-physical system’ and ‘system design approaches’ with that of ‘design principles’. This linking allows us to explore and understand causalities with regards to the aforementioned knowledge gap. The new two knowledge domains, namely ‘design supporting tools’ and ‘system-level features’ domains have been included in the study because of our intention to provide a theoretical framework and technical solution for supporting pre-embodiment design of cyber-physical systems.

The nature of cyber-physical systems was included in the study with the intent of collect-ing information about the characteristics of CPSs, which differentiate them from other technical systems and traditional products. The conducted study has explored what these characteristics are, and why and how these affect the conceptualization, embodiment and

(29)

impleme scope of durables evidence the desig also supp those ava tools in o The study approach for desig approach concentra and com identified character was that tectural d third dom principle The fourt ered the available were take The stud initially c Some of system-le design. T compone Conseque as a possi As an im studied a a differe discussed feature t theory-in methodo Figure 2-1 ntations of our investi (CPCDs). T s that there gn challeng ported our ailable in th order to be a y of the sec hes and to i gning CPSs. hes dedicat ate on CPC putational d. Through ristics of CP none of th design and main of stu level. th studied tools that in web pag en into con y identified created for them were evel design The results ent design d ently, a (ne ible solutio mplication, t as well as so nt objectiv d in a hug echnology ntensivenes ological sym Domains CPSs. In th gation to a he study o e are intrins ges of tradi assumptio he form of C able to han cond doma identify mis . We introd ted to desig CDs. In the approache a critical co PCDs, confl e studied t d functiona udy explor knowledge have been ges and rep nsideration d various f compone e based on , but misse of this ana does not su ew) concept on for multi-the past an ome conce ve. The tra ge number has not re s, strong n mbiosis. In of study e very wide a specific fa f the literat sic differen itional cons on that CPS CAD/E syste ndle the hig

ain was inte smatches c duced a sc

ign for cust

course of es of tradit omparative icting situa traditional d l customiza red causalit e domain w document positories. B if they claim functional c nt design, high level d the neces alysis showe upport syst t of system -level syste nd current ptual frame aditional a r of public eceived en need for a addition to e field of CP mily of CPS ture associa ces betwee sumer dura S designers ems or logi h level com ended to e caused by a coping here tomization. the literatu tional desig study of th ations were design app ation of CP ties of the was suppor ted in the l Both acade med to sup categories but insuff abstractio ssary functi ed that ext em-level de m-level featu m design a developme eworks in w pproaches cations. Ho ough atten multi-disc o the spec PS applicati Ss, namely ated with t en the desig ables. This s need diffe cs/analytic mplexity and xplore and applying tra e by limitin This was in ure study, a gn for mas he specified e identified. proaches co PCDs witho identified rting tools literature, a emic implem pport mains of the too ficiently up n and simp ional capab tension of t esign of CP ure technol and custom ents of feat which it can of feature wever, ela ntion so fa ciplinarily k ification of ons, we nar to cyber-ph this family o gn challeng part of the erent suppo s-based sys d heteroge d evaluate t aditional de ng our atte n line with all possible s customiz d design app . The reaso ould be app out a major mismatch for CPS de as well as t ments and c stream CPS ols. Some o pgraded fo plification, a bilities for m the tools c PSs properly ogy has be ization. ture techno n be re-ope e-based de boration o ar. The reas knowledge f the whole rrowed dow hysical con of CPSs pro ges of CPCD e literature orting tool stem archit eneity of CP the recent d esign appro ention to d our endea e methodo zation have proaches a n of the co plied in the r adaptatio es in the d esign. We c hose which commercia design act of the tools r system d and intend multi-level s urrently us y and suffic een hypoth ology have erationalize esign have of a system sons are su integration e of system wn the sumer ovided Ds and study s than tecting Ss. design oaches design avor to logical e been nd the onflicts e archi-n. The design consid-h were al tools tivities. s were design. ded for system sed for ciently. esized e been d with been m-level uch as n, and m-level

System-level feature-based modeling of cyber-physical systems

Delft University of Technology

System-level feature-based modeling of cyber-physical systems

A theoretical framework and methodological fundamentals

System-level feature-based modeling of

cyber-physical systems

A theoretical framework and methodological fundamentals

Proefschrift

ter verkrijging van de graad van doctor

aan de Technische Universiteit Delft,

op gezag van de Rector Magnificus Prof.ir. K.C.A.M. Luyben;

voorzitter van het College voor Promoties,

in het openbaar te verdedigen op Woensdag 26 April 2017 om

10:00 uur

Door

Shahab POURTALEBI

Master of Science in Industrial Design

This diss

promoto

Compos

Indepen

System

A theor

Keywor

ISBN/EA

An elect

http://r

Copyrig

shahab

sertation h

or: Prof. d

sition of

ndent me

m-level f

retical fra

rds:

AN:

978-tronic ve

repository

ght © 201

b60p@gm

has been

dr. I. Horv

the docto

embers:

feature-mework a

94-028-0

rsion of t

y.tudelft.n

17 by S. P

mail.com

approved

áth

oral comm

based m

and meth

0621-2

this disse

nl/

Pourtaleb

d by the

mittee:

modeling

hodologic

rtation is

bi

g of cybe

cal fundam

available

er-physi

mentals

e at

cal syste

ems

TABLE OF CONTENTS

1.1

PROLIFER

INT

1

RATION OF