International journal of computer integrated manufacturing , tập 23, số 12, 2010

It is to be noted that emphasis is laid on the semantics of the internal profile of the rotational product, which are then used to drive the rotational core insert representa-tion knowled

International Journal of COMPUTER INTEGRATED MANUFACTURING

EDITOR-IN-CHIEFStephen T NewmanDepartment of Mechanical Engineering

University of Bath, Bath BA2 7AY, UK e-mail: IJCIMeditor@bath.ac.uk

INTERNATIONAL EDITORIAL BOARD

EDITOR: ASIA (PACIFIC)

EDITOR: NORTH AMERICA

Paul G Ranky

Department of Industrial and

Manufacturing Engineering

New Jersey Institute of Technology

University Heights, Newark NJ 07102, USA

EDITOR: NORTH AMERICAPaul Kenneth WrightDepartment of Mechanical Engineering University of California, 5133 Etcheverry

Hall Berkeley, CA 94720-1740, USA e-mail: pwright@robocop.berkeley.edu EDITOR: EUROPEFranc¸ois Vernadat

IT & Telecommunications Division European Court of Auditors

12, Rue Alcide de Gasperi

1615 Luxembourg e-mail: Francois.Vernadat@eca.europa.eu CONSULTING EDITORGeorge ChryssolourisLaboratory for Manufacturing Systems, Department of Mechanical Engineering and Aeronautics, University of Patras, Patras 26110, Greece

National Tsing Hua University, Taiwan

M Jeng National Taiwan Ocean University, Taiwan

A Jones National Institute of Standards

& Technology, USA

S Joshi Pennsylvania State University, USA

F Jovane Politecnico di Milano, Italy

H Katayama Waseda University, Japan

K Kosanke CIMOSA, Germany

A Kusiak University of Iowa, USA

A Molina ITESM Monterrey, Mexico

L B Newnes University of Bath, UK

S Y Nof Purdue University, USA

P O’Grady University of Iowa, USA

P Pokorny Technical University of Liberec, Czech Republic

P Rogers University of Calgary, Canada

I Sabuncuoglu Bilkent University, Turkey

M K Tiwari Indian Institute of Technology, Kharagpur, India

R Uzsoy Purdue University, USA

H Van Brussel Katholieke Universiteit TE Leuven, Belgium

H Van Dyke Parunak ERIM Center for Electronic Commerce, USA

J Vancza MTA Sztaki, Hungary

R Veeramani University of Wisconsin-Madison, USA

L Wang University of Sko¨vde, Sweden H.-J Warnecke

Fraunhofer-IPA, Germany

R H Weston Loughborough University, UK

X Xu University of Auckland, New Zealand

Towards expressive ontology-based approaches to manufacturing knowledge

representation and sharingNitishal Chungooraa*, Osiris Canciglieri Jr.b and R.I.M Youngaa

Wolfson School of Mechanical and Manufacturing Engineering, Loughborough University, Loughborough, UK;bLaboratory ofAutomation and Systems (LAS), Pontifical Catholic University of Parana´ (PUCPR), Rua Imaculada Conceic¸a˜o, 1155 Prado

Velho, Curitiba, PR, Brazil, CEP 80215-030(Received 25 April 2010; final version received 24 August 2010)The present capability that ontological approaches offer to formally represent and share manufacturing knowledge

is dependent on the choice of ontological formalism Currently, there exists a spectrum of these formalisms, which isbeing subjectively exploited across multiple domains in design and manufacture Hence, there is an importantprerequisite to achieve an understanding of which family of formalism strictly enables the expressive capture ofsemantics to progress towards meaningful information and viable knowledge sharing This article analyses therelative strengths and weaknesses in employing a ‘lightweight’ ontology versus a ‘heavyweight’ version of theontology to represent and share knowledge between multiple domains in injection moulding design andmanufacture A pertinent direction, from an ontology perspective, is then exposed as a prescription for theimproved capture and dissemination of formal semantics, to support multi-domain knowledge sharing

Keywords: design and manufacture; lightweight ontology; heavyweight ontology; semantics; knowledge sharing;injection moulding

1 Introduction

Ontological approaches are nowadays increasingly

being applied to support the formal capture and

sharing of the meaning and intent (i.e semantics) of

design and manufacture concepts For a particular

domain, the representation of the required semantics is

held in an ontology and the knowledge base (KB)

deployed from the ontology is used to populate

knowledge which should consistently derive from the

semantic structures within the ontology Represented

knowledge in a KB provides useful support for key

engineering decisions, for example the ways in which a

designer’s intent in the design domain could affect the

selection of manufacturing processes in the

manufac-turing domain Thus, expressive manufacmanufac-turing

knowl-edge refers to populated knowlknowl-edge in a KB, based on

the unambiguous definition of semantics structures,

which carry enriched formal meaning

Unfortunately at present, the seamless exchange of

design and manufacture semantics for knowledge

sharing is still not achievable as a result of domain

models that do not carry sufficiently expressive

semantics This is because there are currently several

ontological formalisms, of varying expressiveness (Ray

2004) and system interaction capabilities, which do not

all necessarily address the knowledge capture and

sharing needs in product design and manufacture.Consequently, there exists an ongoing requirement torefine the understanding of the level of logicalexpressiveness capable of semantically structuring themeaning of product lifecycle concepts (Young et al

2009, Chungoora 2010)

This article investigates the capture and system sharing of ontology-based knowledge using thebasis of two broad categories of ontological formal-isms, notably ‘lightweight’ and ‘heavyweight’ ap-proaches (Go´mez-Pe´rez et al 2004), further explained

intra-in the next section By understandintra-ing the implications

of each approach applied to concepts in injectionmoulding design and manufacture, the article con-tributes to a clarification of (1) the ways of expressivelycapturing domain semantics and (2) the mechanismsfor sharing semantics across intra-system domains tosupport engineering decisions

A case study has been devised to expose the relativestrengths and weaknesses between a lightweightontological model and a version of the modelformalised using a heavyweight formalism It hasbeen shown that the existence of an axiom layer inthe heavyweight model is paramount to capturingrigorous semantics and for prompting the potential forknowledge sharing Moreover, certain characteristics

*Corresponding author Email: n.chungoora@lboro.ac.uk

Vol 23, No 12, December 2010, 1059–1070

ISSN 0951-192X print/ISSN 1362-3052 online

Ó 2010 Taylor & Francis

DOI: 10.1080/0951192X.2010.518976

of the lightweight model have proved to be pertinent to

aiding intra-system knowledge sharing Following this

case study, a suitable ontological direction is then

identified, as a benchmark for design and manufacture

domains that intend to exploit expressive semantics

alongside knowledge inference support

2 Lightweight and heavyweight ontological

approaches

2.1 A categorisation based on expressiveness

The requirements and preferences adopted by different

communities have led to the development and

utilisa-tion of various ontological formalisms Ontological

formalisms are essentially formal languages that

support the construction of ontology-based models

and the encoding of the subject matter within these

models Some commonly occurring formalisms are

illustrated in Figure 1, featuring the Unified Modelling

Language (UML 2009), frame-based languages (Wang

et al 2006) and description logic-based languages

(Baader et al 2007) among others To distinguish

families of ontological formalisms, the ontology

community has introduced a categorisation based on

the expressiveness of the subject matter contained

within ontologies, and enabled via the use of

ontolo-gical formalisms

This categorisation involves the notions of

‘light-weight’ and ‘heavy‘light-weight’ ontological approaches,

which primarily differ in the degree of formality and

granularity with which they can represent the same

knowledge (Go´mez-Pe´rez et al 2004, Casely-Hayford

2005) Lightweight models predominantly consist of a

taxonomy of concepts, with simple relationships

established among these concepts and very basic

constraints over the meaning of the ontological terms

On the other hand, heavyweight models, in addition to

having the lightweight structures, are accompanied by

a rich set of formal axioms that constrain theinterpretation of ontological terms In Figure 1,UML and Frames used on its own are examples oflightweight ontological approaches (A), while DL andframes with a first-order logic constraint language areexamples of heavyweight ontological approaches (B)

In the field of manufacturing engineering research,both lightweight and heavyweight methods have beenused for the formalisation of domain models (ISO

18629 2005, Patil et al 2005, Kim et al 2006, Lin andHarding 2007) It is clear, from the extent of theexploited lightweight and heavyweight ontologicalapproaches, that there is currently no discernableconsensus on a preferred ontological direction This islargely because of the ongoing need to establish thesuitability of these approaches to meet the semanticand knowledge sharing requirements of design andmanufacture Hence, the aim of assessing the benefitsand limitations of both approaches becomes a key steptowards identifying the essential elements to progresstowards expressive ontology-based approaches

2.2 Multi-domain knowledge representation andsharing using lightweight and heavyweight approachesThe methodology to achieve the previously mentionedaim is identified in Figure 2 Emphasis is placed onmulti-domain knowledge representation and intra-system knowledge sharing in the context of injectionmoulding The methodology involves considering asimple consumer product concept, namely a rotationalcontainer (C), as shown in Figure 2 Using bothlightweight UML and heavyweight Frames with a first-order logic constraint language, the product represen-tation is first to be captured in the mouldabilitydomain (D), by using the semantic structures sup-ported in both methods

Then, populated knowledge from the mouldabilitydomain is to be shared with the mould design domain(E) for obtaining a representation of the mouldproduct model knowledge Following this stage, themould product model knowledge from the moulddesign domain is to be shared with the mould-manufacturing domain (F) to capture the manufactur-ing representation knowledge for the mould Thesharing process between domains is to be achieved byusing the adequate translation/mapping mechanismsaccommodated in both ontology-based approaches

A number of reasons justify the selection of UMLand Frames with a first-order logic constraint lan-guage, as the preferred lightweight and heavyweightontological formalisms respectively In the first place, arange of lightweight information models has exploitedUML for multi-viewpoint modelling applied to design

Figure 1 Examples of lightweight and heavyweight

ontological approaches

realisation stages (Tam et al 2000, Kugathasan and

McMahon 2001, Canciglieri and Young 2009) Thus,

by performing an assessment of UML against the

exposed methodology, it becomes possible to provide

an appreciation of one extensively used lightweight

formalism

Frames with a first-order logic constraint language

as heavyweight formalism presents characteristics that

overlap with a range of other heavyweight formalisms

This explains its suitability for this investigation For

example the formalism bears several structural

simila-rities to description logic-based languages, which have

witnessed unprecedented relevance in product design

ontologies (AIM@SHAPE 2004, Lukibanov 2005)

Furthermore, the chosen heavyweight formalism holds

key commonalities with the common logic interchange

format (ISO/IEC 24707 2007), which has been used to

encode the process specification language (PSL)

ontology (ISO 18629 2005)

3 Case study

3.1 Overview of case study

Figure 3 identifies the case study scenario for the

analysis of the selected lightweight and heavyweight

ontological approaches, to support multi-domain

knowledge representation and sharing This scenario

provides a more detailed view on the use of themethodology portrayed in Figure 2 Based on Figure 3,the study concentrates on the analysis of a UML multi-domain injection moulding model against a similarmodel formalised in Frames with a first-order logicconstraint language A detailed understanding behindthe UML development of the multiple viewpointdomains can be found in an earlier manuscript(Canciglieri and Young 2003)

For implementation purposes, the KnowledgeEngineering Methodology, prescribed by Noy andMcGuinness (2001), has been adopted during ontologydevelopment An appropriate UML tool has beenutilised to formalise the lightweight UML model.Furthermore, the formalism Frames with the Prote´ge´Axiom Language (PAL), as first-order logic constraintlanguage, has been used in the Prote´ge´ Frames 3.4ontology editor (Prote´ge´ 2009) for representing theheavyweight model

An important facet of the case study is related tothe formal representation of the semantics of multi-domain injection moulding and the correspondingpopulated knowledge This involves the following: the explicit representation of the mouldability,mould design and mould manufacturing domainsemantics

Figure 2 Methodology for multi-domain knowledge representation and sharing

representing knowledge that is either common

across domains or needs to be translated/mapped

to a different domain

capturing logical pre-conditions that exist in one

domain that can drive the translation/mapping

of the appropriate knowledge

The representation of the rotational product in the

mouldability domain is partly comprised of internal

and external profiles (G) that pertain to primary and

transition features that form the wall of the product

The dimensional knowledge captured in these profiles

is to be shared with the mould design domain (H) It is

to be noted that emphasis is laid on the semantics of

the internal profile of the rotational product, which are

then used to drive the rotational core insert

representa-tion knowledge (I) in the mould design domain In this

case, the mould design domain has been referred to as

the ‘rotational core design domain’ to clarify that the

intended representation is for the rotational core

component of the mould

The knowledge shared from the mouldability

domain to the rotational core design domain, is then

used to disseminate additional knowledge to the

rotational core-manufacturing domain (J) (i.e the

mould manufacturing domain) The rotational

core-manufacturing domain represents the semantics of therotational core insert from a machining viewpoint, forexample in terms of the types of machining featuresthat the rotational core insert holds in the manufactur-ing domain (K)

3.2 Lightweight ontology-based modelThe lightweight UML model has been previouslydocumented (Canciglieri and Young 2009) and therefore,this section concentrates on the most relevant strengthsand weaknesses carried by the lightweight UMLapproach Figure 4 provides a broad understanding ofthe implementation of the lightweight ontology-basedmodel In the model, UML class diagrams (L) have beenexploited to represent the necessary concepts and, tosome extent, the basic semantics of each domain TheseUML class representations capture domain concepts inthe form of classes and arrange these classes according to

a taxonomy Relations with cardinality information (M)are used to formulate the key associations that holdbetween different classes

The identification, retrieval and sharing of lated knowledge from the mouldability domain to therotational core design domain and from the designdomain to the manufacturing domain is formalised

popu-Figure 3 Case study scenario

through UML activity diagrams (N) These activity

diagrams enable the user to create a set of instructions

on how to translate the required attributes and

knowledge from one domain to another

3.2.1 Strengths and weaknesses

UML class diagrams provide a convenient way to

design ontologies, because they support a fairly

rich set of graphical constructs This can be a

particularly useful means of reusing

platform-independent ontologies prior to their

implementa-tion in the required ontology applicaimplementa-tions

The representation of multi-domain informationstructures is dominated by the use of UML classdiagrams that involve taxonomies of classes andcardinality relationships between classes, whichare fundamental to any ontology From thelightweight model explored, it has been possible

to exploit UML class diagrams to capturecommon information content across domains There are two main ways in which classes areallowed to carry some semantics namely (1)through the specification of traits that the classespossess (i.e attributes), and (2) by specifyingbinary relations that hold between pairs of classes

Figure 4 Using UML class and activity diagrams in the lightweight model

Classes, attributes and relations in UML hold

textual descriptions rather than semantic

defini-tions Consequently, domain concepts can only be

meaningfully interpreted if the implied semantics

of these concepts are understood by the user

UML activity diagrams allow

translation/map-ping knowledge to be captured and aid, at a system

development level, to automatically perform

information sharing procedures from one domain

to another For example, it is possible to trigger the

automatic assertion of attributes and dimensional

knowledge from the mouldability domain to the

rotational core design domain However, in the

experiment, because UML activity diagrams

depend on UML class diagrams, this implies that

translation procedures are dependent on the terms

carried by domain concepts rather than the

semantics of these concepts

Although a low level of computational

inter-pretation can be captured in UML classes purely

associated to variations in class names, it is not

fully possible to embed pre-conditional

knowl-edge and intent For example in Figure 4, the

class name ‘Rot_Wall_Par_Part_Line’ (O) in the

mouldability domain is used to imply a

rota-tional primary feature, which is positioned

parallel to a parting line configuration (P)

However, the condition for parallelism to a

parting line cannot be formally stated in UML

3.3 Heavyweight ontology-based model

The heavyweight ontological exploration using

Frames with the PAL differs both in the degree of

formality and granularity when compared with the

lightweight approach Figure 5 depicts the heavyweight

ontological structures used to model multi-domainsemantics and to identify sharable knowledge betweendomains

In the heavyweight model, ontological structuresconsist of taxonomies of classes, relations and func-tions, accompanied by a rigorous logic-based axiomlayer as shown in Figure 5 (Q) This layer is responsiblefor supporting the meaning of concepts in computa-tional form The axiom layer is built on top of the basicontological structures and consists of integrity con-straints and mapping rules, which are both written inPAL This constraint language accommodates first-order semantics, thereby providing considerable flex-ibility in specifying the conditions for semanticconformance and knowledge sharing Integrity con-straints are logical restrictions that help to ensure thesemantic integrity within the injection mouldingdomains identified in Figure 3, while mapping rulesare logical conditions that help to identify potentialknowledge that could be shared from one domain tothe other

Figure 6 provides a screen shot of the ‘mouldabilitydomain’ (R) class taxonomy in the class browser,which at first glance is very similar to the classtaxonomy from the lightweight UML class model.Other abstract classes are present namely ‘rotationalcore design domain’ (S) and ‘rotational core manu-facturing domain’ (T), which contain the informationstructures for the rotational core insert design andmanufacture, respectively The abstract class ‘Com-mon Semantics’ (U) regroups reusable behavioursacross domains, such as the notions of ‘point’, ‘axis’,

‘length measure’ and ‘dimensional tolerance’ amongothers

An instance of the class ‘rotational mouldabilityproduct’ (V) is shown in the instance browser

Figure 5 Heavyweight ontological structures

Captured semantics for one specific instance of

‘rotational mouldability product’, named ‘Product 1

-Rotational Container’ (W), can be identified in

Figure 6 These semantic structures involve, for

example the list of point profiles aggregated under

the relations ‘holds_internal_profile’ (X) and

‘hold-s_external_profile’ (Y) and the list of primary and

transition features aggregated under the binary

rela-tion ‘holds_feature’ (Z)

3.3.1 Integrity constraints

From an ontology formalisation viewpoint, PAL is

used for model checking This implies that integrity

constraints act as semantic prescriptions to ensure that

populated knowledge in the KB conforms to the

semantics expressed in the heavyweight model To

verify whether asserted knowledge violates or

con-forms to semantics, integrity constraints can be

processed and a number of results are retained in the

event that these constraints have been infringed In

other words, integrity constraints contribute to the

semantic integrity and enrichment of the KB

To account for the semantic needs of the

heavy-weight model, integrity constraints have been written

for the multiple domains under consideration Over 30

integrity constraints including both simple and

com-plex ones have been modelled for all three domains

The expression listed next gives an example of a simple

integrity constraint in the mouldability domain to

ensure that instances of the class ‘rotational ability product’ (see Figure 6 (V)) are only allowed tohold one axis of rotation

mould-(defrange ?product :FRAME ‘Rotational ability Product’)

Mould-(forall ?product(¼ (number-of-slot-values holds_axis ?product) 1))

If, for example an instance of ‘Rotational ability Product’ is asserted as having more than oneaxis in the KB, then an execution of the PALconstraint would show that this instance is violatingthe fundamental semantics that a rotational mould-ability product must always hold one axis Figure 7illustrates the result of querying an integrity constraintbased on an incorrectly populated knowledge element.The instance ‘Product 1 - Rotational Container’ (W) isshown to be violating the integrity constraint at querytime as a result of an additional ‘Probe InconsistentAxis’ (A1) having been assigned The identification ofinconsistent knowledge, like the one shown in Figure 7,provides a useful way of prompting the user to rectifythe incorrect assertions

Mould-An example of a more complex integrity constraint

in the mouldability domain is shown in Figure 8 Theaxiom captures the relevant logical pre-conditions toensure the correct specification of parting line features(see Figure 4 (P)), by using the appropriate formalisedstatement (B1) In the expression, the accurate

Figure 6 Capturing the semantics of the mouldability domain

definition of a ‘Parting Line Feature’ (C1) is captured

based on the known existence of some defined ‘Primary

Feature’ (D1) A similar understanding has been

followed for the specification of other simple andcomplex integrity constraints required for the moulddesign and mould-manufacturing domains

Figure 8 Example of a complex integrity constraint

Figure 7 Reporting an integrity constraint violation

Once the mouldability domain has been modelled

with its required instances populated, integrity

con-straints for that specific domain are processed to

ensure that the asserted knowledge concords with

semantics If a constraint is violated this implies

that the query response, obtained from running

the constraint, points to inconsistent knowledge The

consequence of a visible inconsistency prompts

the user to modify and/or assert correct knowledge

The process of checking integrity constraints is iterated

until there are no violated conditions, thereby ensuring

the completeness of domain semantics

3.3.2 Mapping rules

The next stage following model checking of the

mouldability domain involves the execution of

map-ping rules A first set of mapmap-ping rules is run with the

intention of identifying sharable knowledge that needs

to be communicated from the mouldability domain to

the rotational core design domain Mapping rules are

written in the same way as integrity constraints and are

very similar in terms of complexity The main

difference between the two lies in the specification of

existential quantifiers (i.e the first-order logic directive

called ‘exist’) in the consequent of the mapping rules

An example of a mapping rule is informally quoted

next, together with an exemplified understanding of the

implications of the mapping rule as shown in Figure 9

‘Rotational core perpendicular straight line(s) must

be specified’ informally says that: If a parallel parting

line primary feature (e.g (E1)) of a rotational

mouldability product has two points ?p1 and ?p2

that describe the internal profile of the product, such

that only the z coordinates of the two points are

different while the x and y coordinates are the same,

then there should exist a core perpendicular straight

line (e.g (F1)) that meets the two points in the

rotational core design domain

In simpler words, logical semantic conditionsarising in the mouldability domain imply the existence

of similar, modified or different knowledge elements inthe rotational core design domain After sharableknowledge is processed on running the first set ofmapping rules, the user then manually creates theidentified knowledge for the rotational core designdomain Once this stage is performed, integrityconstraints for the rotational core design domain areexecuted to ensure the consistency of the new knowl-edge input

The next stage of knowledge sharing involvesdiscovering mappings from the rotational core designdomain to the rotational core-manufacturing domain

An informally expressed example of a mapping rule inthis case is listed next, together with its correspondingexplanatory diagram in Figure 10

‘Horizontal turning feature(s) must be specified’informally says that: If a straight line (e.g (G1)) thatdefines the core insert for a rotational mouldabilityproduct has two points ?p1 and ?p2, such that only the

xcoordinates of the two points are different while the zand y coordinates are the same, then there should exist

a horizontal turning feature profile (e.g (H1)) thatmeets the two points in the rotational core-manufac-turing domain

Results from similar mapping rules help identifynew knowledge required for the rotational core-manufacturing domain based on the knowledgeelements found in the rotational core design domain.The user asserts the identified sharable knowledge inthe rotational core-manufacturing domain and ascer-tains that the knowledge input is consistent withdomain semantics by executing the integrity con-straints for the manufacturing domain

Figure 11 illustrates the results of processing twomapping rules for each domain-to-domain sharableknowledge identification process Overall, 18 complexmapping rules have been explored in the heavyweight

Figure 9 Example of a mapping rule for sharing between

the mouldability and rotational core design domains

Figure 10 Example of a mapping rule for sharing betweenthe rotational core design and manufacturing domains

experiment to obtain, from the product representation

in the mouldability domain, the accurate

representa-tion of the rotarepresenta-tional core insert in both the mould

design and manufacturing domains

In the sample results in Figure 11, it can be seen

that sharable knowledge has been inferred and consists

of (1) hole features from the mouldability domain that

can be directly shared with the rotational core design

domain (I1), (2) product geometry-related semantics

(J1) from the mouldability domain that are required in

the rotational core design domain and (3)

geometry-related semantics (K1) of the design rotational core

insert that are sharable with the rotational core

manufacturing domain, to obtain a machining feature

definition for the rotational core insert

3.3.3 Strengths and weaknesses

The fundamental, primarily lightweight,

seman-tic structures of domain concepts can be readily

modelled through the specification of classes and

their taxonomies, accompanied by binary

rela-tions that hold between classes, and funcrela-tions

that act like attributes of classes

The presence of an axiom layer provides the

capability to support the definition of rigorous

semantic structures, which complement the

light-weight structures The axiom layer

accommo-dates a set of integrity constraints and mapping

rules written in the expressive and relatively

flexible PAL, which is first-order logic based

The axiom layer offers the ability to formally

capture logical pre-conditions Although the

integrity constraints that model these

pre-conditions deserve to be carefully written andmay result in an impedance in processing time as

a result of complexity, yet it is seen that suchconstraints explicitly enable the subject matter ofthe heavyweight model to be represented In the heavyweight approach, the use of integrityconstraints has remained rigid In other words,integrity constraints can only be specified todictate the compulsory conformance of popu-lated knowledge in the KB In certain situations,

it could be necessary to also support optionalconformance of knowledge that is left for theuser to decide, as opposed to relying on thesystem This would require the ability to specifyintegrity constraints of lesser ‘strength’, whileremaining traceable by the user

It is possible via the use of mapping rules tomake inferences for aiding the identification ofsharable knowledge that needs to be mappedfrom one domain to another However, in thechosen heavyweight approach, it has not beenpossible to automatically perform the assertion

of new knowledge as this has relied on manualinput, articulated through the processing ofmapping rules This drawback is due to the factthat PAL is essentially used for writing restric-tions on existing knowledge rather than forasserting new knowledge (Prote´ge´ 2009) Furthermore, the heavyweight ontology devel-opment process is absent of the use of suitableontology design schematics, which would serve

as platform-independent model This is becausethe ontology has been directly constructed withinthe implementation environment Hence, thissuggests that the heavyweight model is plat-form-dependent and, therefore, makes the pro-cess of interoperability between differentapplications a potential issue

4 DiscussionsThe case study has documented a set of strengths andweaknesses of lightweight and heavyweight ontologicalapproaches, applied to multi-domain knowledge re-presentation and sharing in injection moulding designand manufacture One of the primary differencesbetween the two approaches lies in the ability for theheavyweight model to accommodate an axiom layersupported by first-order logic semantics The axiomlayer helps express the behaviours and conditions thatprescribe the integrity of populated domain knowledge

in KBs

In UML 2, the object constraint language (OCL)can be used to specify invariant conditions that must

Figure 11 Samples of processed mapping rules for

multi-domain knowledge sharing

hold for the system being modelled (OCL 2006) This

implies that in a similar way to PAL, OCL would

enable the representation of certain logic-based

condi-tions for ensuring that accurate knowledge is

popu-lated However, OCL does not possess the expressive

power of first-order logic For this reason, heavyweight

ontological approaches are favoured from the

perspec-tive of semantic expressiveness and interoperability

Nevertheless, lightweight UML models can still be

effective in collaborative settings, provided the

con-cepts defined in these models are agreed and

understood

From the perspective of automating knowledge

assertion processes, UML activity diagrams hold a

stronger prospect of making the translation / mapping

of knowledge an easier task compared to mapping

rules Thus, it can be extrapolated that by interfacing

UML activity diagram mechanisms with heavyweight

models, it could be possible to achieve an improved

method for automatically asserting new knowledge In

the Prote´ge´ ontology environment, the Jess rule engine

reasoner (Prote´ge´ 2009) is able to interact directly with

populated knowledge and could, therefore, potentially

be exploited to perform the automatic assertion of

inferred knowledge from PAL mapping rules

Additional opportunities exist for extending the

scope of the heavyweight ontology of injection

moulding design and manufacture For example the

mould-manufacturing domain could be broadened to

include manufacturing process sequencing knowledge

This is where, in particular, the PSL ontology would

help formalise the semantics of flow models (Bock and

Gruninger 2005) Such extensions would require a

heavyweight ontological approach that fully supports

more intricate relations and functions, as a result of the

complexity in the semantics of manufacturing process

sequences For meeting this purpose, it would be

required to identify an even more expressive

heavy-weight ontological formalism, because using Frames

with a first-order logic constraint language in Prote´ge´

imposes certain restrictions to using binary relations

and less powerful semantic structures

On the other hand, UML class diagrams as a

conceptual modelling method presents interesting

possibilities as far as platform-independent ontology

design is concerned This is especially because

cur-rently, there is no de facto ontology design schematic

language Therefore, for example, a UML class can be

used to represent a class in a heavyweight ontology and

a UML binary association can directly map to a binary

relation Another example involves higher-arity

rela-tions, such as ternary relarela-tions, that can be represented

in UML by using the construct of n-ary associations

However, more complex heavyweight constructs like

functions of multiple arities and the instantiation of

meta-classes would demand an agreed mode ofexploiting UML class diagrams to avoid ambiguity

5 ConclusionsThe study presented in this work has shown that therecurrently exist a number of benefits and drawbacksrelated to both lightweight and heavyweight ontologi-cal approaches It is understood from this that theprogress towards enabling expressive manufacturingknowledge representation and sharing is bound toenfold the integration of an array of ontology-basedunderstandings and semantic technologies

Hence, the underpinning towards expressive ogy-based approaches firstly requires enabling multipledomains to explicitly represent fundamental semanticstructures These fundamental structures should in-clude the notion of classes and their taxonomies,together with relations and functions, which can bindmore than two classes together At present, theformalism Frames with a first-order logic constraintlanguage is not able to capture the representation ofthese more complex relations and functions, whileUML also falls somehow short of a direct way for sodoing

ontol-Second, integrity constraints need to be formalised

to complement these fundamental structures, therebysemantically enriching ontologies and ensuring thesemantic consistency of populated knowledge It ishighly desirable that integrity constraints be writtenusing an appropriate first-order logic-based language

to impart the required level of logical expressiveness.Third, the process of ontology development should

be accompanied by the provision of appropriateontology design schematics This is an essential stage

to support the platform-independent representationand design of fundamental structures, prior to theirimplementation Currently, frames with a first-orderlogic constraint language in Prote´ge´ is less suited forthe purpose of ontology design due to platform-dependent modelling Conversely, UML offers usefulprospects for ontology design, especially as it may bepossible to utilise UML in a customised way torepresent more complex ontological notions

Finally, the ability to formalise knowledge ence rules, using a suitable first-order logic-basedlanguage, should be provided as a means of derivingnew and expressive knowledge to support engineeringdecisions By providing a logic-based ground forinference rules, it becomes possible to verify derivedknowledge via tractable reasoning procedures Theseprocedures need to be accompanied by the relevanttranslation/mapping mechanisms so as to help performthe automatic assertion of derived multi-domainknowledge

infer-With the introduction of new ontological

formal-isms, notably Common Logic (CL) (ISO/IEC 24707

2007), it is clear that improved capabilities are

emerging to address higher levels of semantic

expres-siveness and interoperability Work is currently

under-way in our laboratory to take this aspect forward

Acknowledgements

The research developed in this article has resulted from a

number of strands of work that have been supported by

different funding agencies In particular, we wish to thank the

EPSRC, who are funding the majority of our work on

‘Interoperable Manufacturing Knowledge Systems’ (IMKS)

under project 253 of the Loughborough University

Innova-tive Manufacturing and Construction Research Centre, and

the Wolfson School of Mechanical and Manufacturing

Engineering of Loughborough University for funding

research studentships

References

AIM@SHAPE, 2004 Advanced and innovative models and

tools for the development of semantic-based systems for

handling, acquiring and processing knowledge embedded in

multidimensional digital objects Available from: http://

www.aimatshape.net/ [Accessed March 2010]

Baader, F., Horrocks, I., and Sattler, U., 2007 Description

logics In: F van Harmelen, V Lifschitz, and B Porter,

eds Handbook of knowledge representation Amsterdam:

Elsevier, 135–179

Bock, C and Gruninger, M., 2005 PSL: a semantic domain

for flow models Software and Systems Modelling

Journal, 4, 209–231

Canciglieri, O.J and Young, R.I.M., 2003 Information

sharing in multi-viewpoint injection moulding design and

manufacturing International Journal of Production

Re-search, 41, 1565–1586

Canciglieri, O.J and Young, R.I.M., 2009 Information

mapping across injection moulding design and

manufac-ture domains International Journal of Production

Re-search DOI: 10.1080/00207540902824974

Casely-Hayford, L., 2005 Environments, methodologies and

languages for supporting users in building a chemical

ontology MSc Dissertation Faculty of Life Sciences,

University of Manchester

Chungoora, N., 2010 A framework to support semantic

interoperability in product design and manufacturePh.D

Thesis Loughborough University

Go´mez-Pe´rez, A., Ferna´ndez-Lo´pez, M., and Corcho, O.,

2004 Ontological engineering: with examples from the

areas of knowledge management, e-commerce and the

semantic web London, UK: Springer-Verlag

ISO 18629, 2005 Industrial automation systems and tion – process specification language (PSL) Geneva,Switzerland: International Organization for Standardiza-tion (ISO)

integra-ISO/IEC 24707, 2007 Information technology – common logic(cl): a framework for a family of logic-based languages.Geneva, Switzerland: International Organization forStandardization (ISO)

Kim, K.-Y., Manley, D.G., and Yang, H., 2006 based assembly design and information sharing forcollaborative product development Computer AidedDesign, 38, 1233–1250

Ontology-Kugathasan, P and McMahon, C., 2001 Multiple viewpointmodels for automotive body-in-white design Interna-tional Journal of Production Research, 39 (8), 1698–1705.Lin, H.K and Harding, J.A., 2007 A manufacturingengineering ontology model on the semantic web forinter-enterprise collaboration Computers in Industry, 58(5), 428–437

Lukibanov, O., 2005 Use of ontologies to support designactivities at DaimlerChrysler Proceedings of the 8thInternational Prote´ge´ Conference,Madrid, Spain.Noy, N.F and McGuinness, D.L., 2001 Ontology develop-ment 101: a guide to creating your first ontology.Knowledge Systems Laboratory, Stanford University.Object Constraint Language (OCL), 2006 OMG availablespecification version 2.0 Available from: http://www.om-g.org/technology/documents/formal/ocl.htm [AccessedMarch 2010]

Patil, L., Dutta, D., and Sriram, R., 2005 Ontology-basedexchange of product data semantics IEEE Transactions

on Automation Science and Engineering, 2 (3), 213–225.Prote´ge´, 2009 Ontology editor and knowledge acquisitionsystem Available from: http://www.protege.stanford.edu[Accessed November 2009]

Ray, S.R., 2004 NIST’s semantic approach to standards andinteroperability PowerPoint presentation Availablefrom: http://ontolog.cim3.net/file/resource/presentation/NIST_Semantics–SteveRay_20040212a.ppt [AccessedNovember 2009]

Tam, S., et al., 2000 An object-based process planning andscheduling model in a product design environment.Logistics Information Management, 13 (4), 191–200.Unified Modelling Language (UML) Available from: http://www.uml.org/ [Accessed November 2009]

Wang, H.H., et al., 2006 Frames and OWL side by side.Proceedings of the 9th International Prote´ge´ Conference,Stanford, CA

Young, R.I.M., et al., 2009 Enabling interoperable facturing knowledge sharing in PLM In: Proceedings ofthe 6th international product lifecycle management con-ference,University of Bath, Bath, UK

Automatic inspection of turbine blades using 5-axis coordinate measurement machine

Hui-Chin Changa* and Alan C Linba

Department of Mechanical Engineering, De Lin Institute of Technology, No 1, Lane 380, Ching-Yun Road, Tu-Cheng,Taipei, Taiwan, Republic of China;bDepartment of Mechanical Engineering, National Taiwan University of Science and

Technology, 43 Keelung Road, Section 4 Taipei, Taiwan, Republic of China(Received 25 October 2009; final version received 17 September 2010)The complex geometric shape of turbine blades not only causes difficulties in fabrication but also makes it difficult toexamine the part’s precision when using a traditional 3-axis coordinate measurement machine (CMM) Generallyspeaking, one has to use a multi-axis CMM and follow the appropriately planned measuring paths to accomplish thetask of precision measurement of the machined part This article discusses the methodology of using a 3D geometricmodel of a turbine blade as the basis to generate interference-free measuring paths that are suitable for implementing

on a CMM with three translational axes and two rotational axes Through the use of the methodology proposed inthis research, the goal of multi-axis precision inspection for any geometric design of turbine blade can be achieved.Keywords: automated inspection; CMM; collision-free; turbine blades; 5-axis measurement

1 Introduction

Figure 1 shows the geometric shape of an axial wheel

with a pattern of turbine blade, which is one of the

indispensable components that are found in large

quantities in the aerospace and power industries The

complex geometric shape and the limited space

between two neighbouring blades hinder the

applica-tion of tradiapplica-tional techniques to inspect its precision

In the process of measurement, an unexpected collision

caused by the measuring probe will break off the

measuring process or the measuring equipment itself to

be damaged Therefore, the principle function of path

planning for coordinate measurement of a turbine

blade is to generate numerous orientations and

positions of the probe stylus, which are free of collision

during the measuring process The traditional way of

planning measuring paths mostly relies on manual

editing and online coordinate measurement machine

(CMM) teaching, which is a time-consuming, highly

skilled and error-prone task Generating an

auto-mated, collision-free measuring path thus becomes a

crucial issue for improving the quality of the

measur-ing, as well as reducing the production lead-time for

high-precision turbine blades

For prismatic parts with planar surfaces, the

concept of ‘ray tracing’ can be used to inspect the

collision between the start point and the target point of

the measuring path; if there is a collision, the algorithm

works through the topological structure of the part

and selects the midpoint of the edge, shared by the face

with which the path collides and the adjacent facenearest to the target point, as the next probe point.This procedure is followed till the target point isreached (Lin and Murugappan 1998, 1999) Themethod is valid only for planar surfaces but not forsculptured surfaces Furthermore, there is an approach

to using the geometric features set up beforehand togenerate measuring paths (Hermann 1997) Thisapproach can generate collision-free measuring pathsfor objects with non-prismatic surfaces, such asspherical surfaces, but for complicated sculptured-surfaces, its feasibility cannot be verified As for thegeneration of collision-free measuring paths forcomplicated sculptured-surfaces, Yau and Menq(1995) and Menq and Lai (1992a,b) used 3D CADmodels of object and probe to determine whether anintersection exists between them This approach isaimed at the inspection for objects with complicatedsculptured-surfaces, such as moulds and dies Further-more, various media, such as digital image data(Takeuchi et al 1990) and measured data points inreverse engineering (Gupta and Sagar 1993), have beenstudied in the past to plan appropriate paths for CMM

to undertake the measurement tasks Nevertheless,when it comes to the particular geometric shape ofturbine blades, in addition to their complicatedsculptured-surfaces, there are problems of overlapand undercut between blades

Chang and Lin (2005) using a 3-axis CMMcooperate with a 2-axis automatic fixture to develop

*Corresponding author Email: chang.hcjang@gmail.com

Vol 23, No 12, December 2010, 1071–1081

ISSN 0951-192X print/ISSN 1362-3052 online

Ó 2010 Taylor & Francis

DOI: 10.1080/0951192X.2010.527371

the computer-aided measuring system, and the

colli-sion-free measuring path planning Heo et al (2008)

based on the ruled line information of a CAD database

for impeller blades, they projected hub and blade curve

on the XY plane, and to find the appropriate probe

postures according to the distribution of these

pro-jected curves that enable the probe to gauge the

inspection points without any collision between the

probe and the blade surfaces In the meantime, they

partitioned the target blade surface into several

UMRs, which keeps the same probe approach vector

in each UMR so as not to change the probe

orientation Then, the probe can be taught quite

simply in advance based on the ruled line information

of blade surface

In addition, there are several studies of 5-axis NC

machining related to this topic, e.g Balasubramaniam

et al (2000, 2003) developed methods for 5-axis tool

positioning that account for accessibility of the tool

using visibility maps of the triangulated data though

the tool positions were not optimised for the highest

material removal rate Gray et al (2005) developed the

5-axis AIM, which is based on the fact that the widest

machined strip width is cut when the tool is tilted along

the feed direction The concept of AIM is to tilt the

tool such that the tool axis remains in the tilting plane

and the forward bottom point of the tool remains in

tangential contact with the ccp (cutter contact point)

until a second contact point on the surface is touched

This will give the smallest tilt angle and thus, the

largest effective radius resulting in the widest machined

strip width for the given feed direction at the ccp

without gouging the surface

Considering the above problems as found in the

literature, this article addresses the methodology to

automate the planning of measuring paths of the 5-axis

CMM to suit the particular geometric shape of turbine

blades It also discusses the implementation of the

proposed methodology to develop a computer-aided

measuring system for an automatic multi-axis

preci-sion inspection for turbine blades The application of

the system significantly reduces the time spent on

manually calculating and editing the collision-free

measuring positions of the probe head, a bonus whencompeting in the global marketplace

2 Configuration of measurementBesides the X-, Y-, and Z-axis, a typical 5-axis CMM isequipped with a probe that provides motions in the A-and B-axis, as illustrated in Figure 2 In this article, theorientation of the stylus is represented by (a, b), where

a and b are inclination angle and rotation angle,respectively, and are defined as the rotational angles ofthe stylus for the A- and B-axis The initial orientation

of the A-axis of the stylus is set to be parallel to theX-axis, i.e the initial value of the inclination angle a is908

Every turbine blade in the rotor of a turbo-machinepresents an identical geometric shape Therefore, it isonly necessary to take one blade for the planning of themeasuring paths In this article, the Z-axis is set to bealong the rotational axis of the part, and the X–Yplane is in the planar surface that slices the part, asdepicted in Figure 3

3 Establishing the criteria for interference avoidanceTwo types of interference can be found in thecoordinate measurement of a turbine blade: inter-ference between two blades and undercut interference

of a single blade These interferences can be avoidedthrough proper adjustment of rotation angle b andinclination angle a, respectively, as illustrated inFigures 4 and 5 The methods are stated in thefollowing sections

3.1 Avoiding the interference between two bladesFigure 6 shows relevant angles between the probestylus and turbine blades In the figure, d is the angle

Figure 2 Measuring probe of a 5-axis CMM

Figure 1 Geometric shape of a workpiece with 19 pieces of

turbine blade

between two adjacent blades, f is the angle from the

X-axis to the line passing through the measuring point

P(Px, Py, Pz), and y is the angle that the current

measuring point rotates around the Z-axis untiltouching the next blade The above three angles can

be formulated into the following equations:

measurement

Figure 4 (A) Interference between two turbine blades, (B) adjustment of rotation angle b to avoid the interference

Figure 5 (A) Undercut interference of a single turbine blade, (B) adjustment of inclination angle a to avoid the interference

surface that the measuring point locates has to be

taken into account The equations are as follows:

for measuring points locate on the pressure surface:

for measuring points locate on the suction surface:

Figures 6 and 7 illustrate the calculation of angle

b for a measuring point located on the pressure

surface, and Figures 8 and 9 are for a point in the

suction surface In Equations (4) and (5), the

rotation of the stylus by angle b makes the styluslocate on the middle of the safety boundaries, i.e theangle between the stylus and any of the two safetyboundaries is y/2, as shown in any of the above fourfigures This implies that the stylus has to be rotatedfor every single point to be measured, whether acollision happens or not To minimise the totalnumber of stylus rotations while fulfilling therequirement of free-of-collision, the stylus remainsnot-rotated if the difference of angle b of the firstmeasuring point and the second one is smaller than(1/n)6(y/2), where n¼ 1.5 by default in this study,and can be altered by the user

Figure 6 Relevant angles for measuring a point on the

pressure surface and f 5 0

Figure 7 Relevant angles for measuring a point on the

pressure surface and f 0

Figure 8 Relevant angles for measuring a point on thesuction surface and f 5 0

Figure 9 Relevant angles for measuring a point on thesuction surface and f 0

In a typical CMM, the minimal unit of change of

rotation angle of the stylus is 7.58; that is to say, the

angle of stylus orientation is always a multiple of 7.58

If angle b is divided by 7.58, then the result can be

represented by an integer I and a decimal fraction R/

The criteria developed in the earlier section for suction

and pressure surfaces only consider the interference in

the X–Y plane, without addressing the undercut

problem caused by the Z-directional twist of the

turbine blade The front view in Figure 10 depicts an

undercut area of a suction surface

During the probe-in process, undercut interference

occurs if the stylus touches any point on the silhouette

of the blade surface Assuming that the initial

orientation of A-axis of the stylus is set to be parallel

to the X-axis, i.e the initial value of inclination angle a

is 908 To acquire the collision point N(x, y, z) on the

silhouette, the rotation angle b of the stylus must be

found beforehand by employing the equations listed in

the earlier-mentioned section With the current stylus

orientation (a, b), a¼ 908, a straight line L(x, y, z) can

be drawn which passes through the measuring point

P(x, y, z) on the blade surface, the axis of revolution of

the turbine blade, and the silhouette of the blade

surface Collision point N(x, y, z) is thus the point on

the silhouette, as illustrated in Figure 11 The

horizontal distance D(x, y) between points P(x, y, z)

and N(x, y, z) can be calculated by:

Dðx; yÞ ¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiðPx NxÞ2þ ðPy NyÞ2

q

ð8Þ

where Pxis the x coordinate of measuring point P(x, y,

z); same with Py, Pz, Nx, Nyand Nz Height difference

h(z) between the two points is (see Figure 12) asfollows:

The real height difference H(z) has to reflect thediameter d of the ruby ball located at the tip of thestylus:

HðzÞ ¼ Nz Pzþ d=2 for pressure surface ð10ÞHðzÞ ¼ Nz Pz d=2 for suction surface ð11Þ

To avoid the occurrence of undercut interference, theinclination angle a of the stylus has to be adjusted by:

Figure 10 Undercut area of a pressure surface

Figure 11 Top view of collision point N(x, y, z)

Figure 12 Sectional view of undercut interference

3.3 An example

The example shown in Figure 14 is used here to

illustrate how the above formulae work for the

cal-culation of angles a and b In the example, the total

number of turbine blades n is 15, the coordinate of the

measuring point P at the pressure surface is assumed to

be (18.67, 78.06, 78.39), the cross-section thickness t

at point P is 2 mm, and the diameter d of the ruby ball

is 2 mm Equations (1)–(3) are used to find angles d, f

Once angle b is determined, collision point N(x,

y, z) can be found by forming a straight line, whichpasses through the Z-axis, the measuring point P andthe silhouette of the current blade N(x, y, z) is thepoint on the silhouette and its coordinate is found to

be (36.72, 715.21, 77.37) From Equation (8), thehorizontal distance D(x, y) between points P and Ncan be calculated:

Dðx; yÞ ¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiðPx NxÞ2þ ðPy NyÞ2q

¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

½ð18:67  36:72Þ2þ ð8:06 þ 15:21Þ2q

¼ 19:41The height difference H(z) is found by Equation (10):HðzÞ ¼ Nz Pzþ d=2 ¼ 7:37 þ 8:39 þ 2=2 ¼ 2:02

The positive value of H(z) implies that the height of thecollision point N is larger than that of the measuringpoint P In other words, undercut interference occursfor the measurement of point P To eliminate theinterference, the stylus is inclined by angle a andEquations (12) and (13) are used to find its value:

a¼ 90þ o ¼ 90þ tan1ðHðzÞ=Dðx; yÞÞ

¼ 90þ tan1ð2:02=19:41Þ ¼ 95:94:Since the minimal unit of angle change of the stylus

is 7.58, angle a is set to be 97.58

Figure 15 lists the final results of applying theabove procedure to calculate angles a and b for pointsalong the three rings of a turbine blade (20 points ofeach ring)

4 System implementationThe CMM adopted in this research to undertake themeasurement task of turbine blades is MitutoyoKN807 equipped with a measuring probe RenishawPH9A The machine has three translational axes, X-,Y-, and Z-axis, and two rotational axes, A- and B-axis,which are able to rotate by 1058 and 1808, respectively,

at 7.58 per angle increment A total of 720 stylusorientations support the measurements of complex

Figure 13 Adjustment of inclination angle a to avoid

undercut interference

Figure 14 An example for calculation of stylus orientation

surfaces The control of the machine and its probe, as

well as the input and retrieval of the measuring data

are achieved by interface IEEE 488

Regarding the development of a computer system

to implement the aforementioned methods for the

calculation of probe orientations, ACIS by Spatial

Technology, Inc., was used as the geometry kernel, and

Visual Cþþ was adopted as the programming

language Detailed steps and some issues required to

be taken care of are listed below:

(1) First and foremost, the 3D CAD model of a

turbine blade is created using

Pro/ENGI-NEER Points to be measured are generated

on the part model, as pictured in Figures 15(a)

and 16 The coordinates, altogether with the

3D CAD model, become the input data into the

developed computer system Normal vectors of

the measuring points are generated by the

system, and the resultant data are exported in

DMIS format so that the data can be used by a

typical CMM Figure 17 shows a sample of the

exported data

(2) The criteria setup in Section 3 are applied next

to find the stylus orientation (a, b) for every

measuring point These data are then insertedinto the DMIS data

(3) The DMIS data are converted into machinecodes acceptable by Mitutoyo KN807 by thepost processor developed in this research.Figure 18 shows the dialogue box of theDMIS data and CMM codes

(4) Generate collision-free measuring paths andinsert them into the CMM codes

Figure 15 (A) Three rings of measuring points on a turbine blade, (B) angles b and a calculated for the measuring points

Figure 16 Three rings of measuring points on a turbineblade

If the change of stylus orientation is a

prerequi-site for the measurement of a specific point Pi,

then the change of stylus orientation should be

done away from the machined part, in order not

to cause a collision during the change of the

stylus orientation As shown in Figure 19, when

the probe finishes the measurement of point

Pi71, which is a point before Pi, the probe

retracts to point A, moves from point A to B, lifts

from point B to C, changes stylus orientation at

point C, and then resumes its probing action

The probe moves downwards from C to B0 (B0

denotes the new stylus orientation), and finally

reaches approach point P0

iof point Pi nates of points B, B0and C are calculated using

Coordi-the following equations:

Bx¼ 1:5  turbine blade diameter  cos ai1

By ¼ 1:5  turbine blade diameter  sin ai1

ai1¼ tan1

Pi1

B0x¼ 1:5  turbine blade diameter  cos ai

B0y¼ 1:5  turbine blade diameter  sin ai

B0z¼ P0iz

ai¼ tan1

PiyPix



(5) Place the machined turbine blade on the CMMtable and conduct coordinate alignment.Whenplacing the part on the measurement machinetable, one will face the problem of aligning thecoordinate system of the CAD model, (Xcad,

Ycad, Zcad), with that of the measurementmachine (Xcmm, Ycmm, Zcmm) Figure 20 is atypical method of coordinate positioning on aCMM First pick three points with the measur-ing probe to form an X–Y plane Then take twopoints to define the X-axis Finally, obtain therotational axis of the part’s central hole toplace the Z-axis The six degrees of freedom are

Figure 17 Measuring points and normal vectors output in

DMIS data format

Figure 18 Sample of DMIS data and CMM control codes

Figure 19 Travelling path during stylus angle change

thus fixed and the coordinate system is defined.

For the measurement of a turbine blade, the

X–Y plane and the Z-axis for measurement can

easily be defined by plane A and axis O on the

part, as depicted in Figure 21 However, the

X-axis of the part may pass through any point

on the blade surface, and thus the axis cannot

be defined in a straightforward manner The

following steps are used to define the X-axis for

measurement: When placing the turbine blade,

observe in bare eye a minute angle between the

X-axis of the blade surface, Xcad, and the X-axis

of the CMM, Xcmm, as shown in Figure 22

Randomly select a point Pcad from the CAD

model and record its coordinate C and normal

vector K The CMM is operated manually

based on C and K, and the measured

coordi-nate is, presumingly, Pcmm After probe-radius

compensation, compare it with point Pcad to

obtain angle deviation s This angle is used to

correct the coordinate system of the

measure-ment process using the coordinate

transforma-tion functransforma-tion embedded in the CMM Based on

the new coordinate system, the CMM is

operated again manually, based on C and K.The measured coordinate is again comparedwith point Pcad, and the angle deviation is usedfor coordinate correction Repeat the sameprocedure until the difference between Pcadand

Pcmm fits the expectations, as shown in Figure

23 If the difference between Pcmm and Pcadcannot be effectively minimised, an alternativethat changes point Pcad and reinitiates X-axispositioning might resolve the problem

(6) Use the CMM codes generated in Step (4) toconduct the actual coordinate measurementtask Once the measurements of all points arecompleted, conduct probe-radius compensation

by means of the points’ normal vectors Thereal turbine blade surface’s data can thus beproduced

(7) Compare the data of compensated points withthe standard data that were obtained from the3D CAD model, as shown in Figure 24, and theerror in between will appear Figures 25–27 are,respectively, the errors of the X-, Y- and Z-coordinate of the inner ring, middle ring andouter ring

5 ConclusionsTurbine blades are widely used, and are especiallyindispensable in the aerospace industry But, thecomplex surface geometry makes the inspectiondifficult to execute One usually must resort to a 5-axis measuring instrument and spend a long timemanually planning the online measuring paths Thisarticle based on the geometry information of a CADdatabase for turbine blades, and use simple basictrigonometry to calculate and generate of 5-axis

Figure 21 Positioning the XY-plane and Z-axis for

Figure 20 Typical method of coordinate positioning on a

CMM

collision-free paths for the measurement of turbineblades The space between two neighbouring blades isconsidered to find the proper rotation angle b for boththe suction and the pressure surfaces On the otherhand, undercut interference of a single blade is avoided

by the adjustment of the inclination angle a Thesystem implementation was carried out by usingMitutoyo KN807 as the CMM, ACIS as the geometrickernel and Visual Cþþ as the system developmenttool The system automatically generated collision-freemeasuring paths in a short time Successful implemen-tation of measuring machined turbine blades verifiedthe viability of the proposed methodology for theautomatic planning of 5-axis measuring paths

References

Balasubramaniam, M., Sarma, S., and Marciniak, K., 2003.Collision-free finishing toolpaths from visibility data.Computer-Aided Design, 35, 359–374

Balasubramaniam, M., et al., 2000 Generating 5-axis NCroughing paths directly from a tessellated representation.Computer-Aided Design, 32 (4), 261–277

Chang, H.C and Lin, A.C., 2005 Automatic inspection ofturbine blades using a 3-axis CMM together with a 2-axisdividing head International Journal of Advanced Manu-facturing Technology, 26 (7/8), 789–796

Gray, P., Bedi, S., and Ismail, F., 2005 Arc-intersect methodfor 5-axis tool positioning Computer-Aided Design, 35(7), 663–674

Gupta, V.K and Sagar, R., 1993 A PC-based systemintegrating CMM and CAD for automated inspectionand reverse engineering International Journal of Ad-vanced Manufacturing Technology, 9, 306–310

Hermann, G., 1997 Feature-based off-line programming ofcoordinate measuring machines In: INES Proceedings ofthe IEEE international conference on intelligent engineer-ing systems, 15–17 September 1997, Budapest, Hungary.545–548

Heo, E.Y., et al., 2008 Computer-aided measurement planfor an impeller on a coordinate measurement machinewith a rotating and tilting probe Robotics and Computer-Integrated Manufacturing, 24, 788–795

Lin, Y.J and Murugappan, P., 1998 A new algorithm forCAD-directed CMM dimensional inspection Proceed-ings of the IEEE international conference on robotics andautomation, 16–20 May, Leuven, Belgium, 893–898

Figure 24 Measured points and original contour of the

Lin, Y.J and Murugappan, P., 1999 A new algorithm for

determining a collision-free path for a CMM probe

Machine Tools and Manufacture, 39, 1397–1408

Menq, C.H and Lai, G.Y., 1992a An intelligent planning

environment for automated dimensional inspection using

coordinate measuring machines Journal of Engineering

for Industry, 114, 222–230

Menq, C.H and Lai, G.Y., 1992b Automated precision

measurement of surface profile in CAD-directed

ins-pection IEEE transactions on robotics and automation, 8

(2), 268–278

Takeuchi, Y., Shimizu, H., and Mukai, I., 1990 Automaticmeasurement of 3-dimensional coordinate measuringmachine by means of CAD and image data Annals ofthe CIRP, 39 (1), 565–568

Yau, H.Y and Menq, C.H., 1995 Automated CMM pathplanning for dimensional inspection of dies and mouldshaving complex surfaces Machine Tools & Manufacture,

35 (6), 861–876

A wireless sensor network-based approach to large-scale dimensional metrology

Maurizio Galetto*, Luca Mastrogiacomo and Barbara Pralio

DISPEA, Politecnico di Torino, C.so Duca degli Abruzzi 24, 10129 Torino, Italy(Received 16 March 2010; final version received 19 August 2010)

In many branches of industry, dimensional measurements have become an important part of the production cycle, inorder to check product compliance with specifications This task is not trivial especially when dealing with large-scale dimensional measurements: the bigger the measurement dimensions are, the harder is to achieve highaccuracies Nowadays, the problem can be handled using many metrological systems, based on differenttechnologies (e.g optical, mechanical, electromagnetic) Each of these systems is more or less adequate, dependingupon measuring conditions, user’s experience and skill, or other factors such as time, cost, accuracy and portability.This article focuses on a new possible approach to large-scale dimensional metrology based on wireless sensornetworks Advantages and drawbacks of such approach are analysed and deeply discussed Then, the article brieflypresents a recent prototype system – the Mobile Spatial Coordinate-Measuring System (MScMS-II) – which hasbeen developed at the Industrial Metrology and Quality Laboratory of DISPEA – Politecnico di Torino The systemseems to be suitable for performing dimensional measurements of large-size objects (sizes on the order of severalmeters) Owing to its distributed nature, the system – based on a wireless network of optical devices – is portable,fully scalable with respect to dimensions and shapes and easily adaptable to different working environments.Preliminary results of experimental tests, aimed at evaluating system performance as well as research perspectives forfurther improvements, are discussed

Keywords: dimensional metrology; large-scale metrology; large-volume metrology; coordinate-measuring systems;mobile measuring system; wireless sensor networks

1 Introduction

Owing to recent advances in integrated circuits and

radio technologies, the use of distributed sensor

networks is more and more widespread for a variety

of applications, such as, environmental monitoring,

indoor navigation, people and objects tracking,

logis-tics, surveillance, industrial diagnostics and other

activities The facts of being typically composed by

compact, lightweight and cheap devices make the

wireless sensor networks (WSNs) appealing also for

other possible uses Among these, the field of

dimen-sional metrology certainly offers a challenging and

interesting scenario, especially when the dimensions to

be measured are of the order of several metres

(large-scale/large-volume dimensional metrology)

Nowadays, there are different instrumental

solu-tions that allow performing dimensional measurements

of large-size objects In many cases, they are

centra-lised systems where a single hardware unit is

respon-sible for the measurements Often, these instruments

are unwieldy and hardly transportable and they

usually show some coverage problems when dealing

with complex working volumes (Peggs et al 2009)

Recently, a few distributed solutions have beenproposed, in which a set of metrological stationscooperates to measure the geometrical features of anobject These systems generally consist of a central unit

to gather and elaborate data coming from a set ofdistributed sensor devices (Maisano et al 2009) Forthis reason, they cannot be considered as completelydistributed, even if they represent a valid attempttowards scalable and flexible systems

Currently available systems for dimensional trology applications are based on different technolo-gies (e.g optical, mechanical, electromagnetic andinertial), providing for several performance, opera-tional, logistic and economic issues Automationrepresents a key feature in an attempt to perform fastmeasurements with an easy-to-use instrument forpossibly untrained operators (Ganci and Handley1998)

me-This article presents a new concept of large-scaledimensional metrology based on WSNs Althoughmost of the existing systems rely on a centralised unitfor measurement and/or data processing, the novelty

of the proposed approach is based on the distributed

*Corresponding author Email: maurizio.galetto@polito.it

International Journal of Computer Integrated Manufacturing

Vol 23, No 12, December 2010, 1082–1094

ISSN 0951-192X print/ISSN 1362-3052 online

Ó 2010 Taylor & Francis

DOI: 10.1080/0951192X.2010.518322

nature of the measuring technology and its extended

automation capabilities, involving hardware, software

and process issues This concept gets together

flex-ibility and portability, characteristics of a distributed

architecture, with capabilities of coordinated and

cooperative control peculiar to intelligent WSNs The

embedded data elaboration hardware allows to

par-tially share the overall computational load, thus

providing for real-time measurement data processing

The software automation inheres availability of

dimensional data as well as geometric features of the

measured object, besides spatial coordinates of single

points This entails a reduction of required operator

skills, being his/her duties limited to sensor carriage

and/or data acquisition task management In addition,

process automation is implemented as to the setup’s

calibration phase and the data elaboration task

According to its working principles, the system is

able to self-localise and characterise the remote sensing

units, without needing any information concerning

their technical specifications and spatial location/

orientation This aspect gives invaluable flexibility to

the system as different sensor devices, having unknown

characteristics, can be used in the same setup and

contribute to the measurement process

In order to analyse the feasibility of the proposed

approach for performing dimensional measurements

of large-size objects, a prototype – MScMS-II (Mobile

Spatial coordinate Measuring System) – has been

designed and experimentally tested, in terms of sensor

device management, measurement data acquisition

and achievable system performance

In section 2, a background, inhering current

avail-able technologies for large-scale metrology (LSM), their

working principles and configuration layout, is

pre-sented Then the article focuses on the concept of

distributed coordinate-measuring systems, in particular

on the possible use of WSNs to this purpose Advantages

and drawbacks of distributed rather than centralisedsystems are analysed and discussed Section 3 brieflyintroduces the prototype system MScMS-II Prelimin-ary experimental results, aimed at providing a metrolo-gical characterisation of the system, are discussed insection 3 as well Finally, section 4 reports someconcluding remarks and proposes future researchdevelopments

2 Large-scale dimensional metrologyDimensional metrology is the branch of metrology thatdeals with measurements of geometrical features.LSM, in particular, can be defined as ‘the metrology

of objects in which the linear dimensions range fromtens to hundreds of meters’ (Puttock 1978) LSMincludes a wide range of applications, such asdimensional measurements of large structures, mon-itoring of deformations, alignment procedures inmanufacture assembling, in many different industrialsectors, such as aerospace, railway and shipbuilding.The large number of existing metrological solutionscan be classified according to different criteria, includ-ing measuring technologies, working principles, mea-surement procedures and system architectures.Although a wide variety of technologies has beenimplemented for dimensional metrology applications,

at present, optical-based systems are largely used due

to their advantages over the other approaches as tometrological performance and their potentialities forLSM applications A taxonomy of main existing LSMsolutions is reported in Table 1, referring to measure-ment procedures (contact and non-contact instru-ments), working principles (multiple lengths, onelength and two angles and multiple angles) and systemarchitectures (centralised and distributed)

According to measurement procedures, a tion has been proposed in Maisano et al (2009),

classifica-Table 1 Taxonomy of available solutions for metrology applications

SYSTEM ARCHITECTUREMEASUREMENT

TDoF-based systemsTwo angles and one length Laser trackers

Total stations

Hi-BallMulti camera based systems

Two angles and one length Theodolites

Laser radars

Single/stereo camera based systems Multi camera based systems

distinguishing between contact and non-contact

instru-ments Contact systems perform measurements by

touching the workpiece, either through a mechanical

arm or through a portable probe On the contrary,

non-contact systems do not need to touch the object to be

measured

A further categorisation of the available

instru-ments has been reported in Cuypers et al (2009),

referring to their working principles Three groups

have been identified, depending on whether they

perform measurements by using multiple lengths, one

length and two angles or multiple angles The former

method, first used by the well-known

coordinate-measuring machines (CMMs) to evaluate the

posi-tion of its remotely controlled contact probe by

knowing the probe geometry and the distances from

the three reference axes along which it moves, is

common to the systems implementing a

multilatera-tion approach as well More specifically, these

techniques use the known locations of three or

more reference points, and the measured distance

between the point to be localised and each reference

point The unknown coordinates can be found by

solving a nonlinear optimisation problem

(Fran-ceschini et al 2009a) This approach is very similar

to GPS (global positioning system) localisation

principle (Hofmann-Wellenhof et al 1997)

Multi-lateration principles are used by the measurement

systems based on laser interferometers (Cuypers

et al 2009) as well as by those based on ToF

(time of flight) or TDoF (time difference of flight)

(Franceschini et al 2009a) On the other hand, many

coordinate-measuring systems rely on the

determina-tion of one length and two angles These systems,

also called spherical coordinate-measurement

sys-tems, locate a point with reference to a spherical

coordinate system Examples of such systems are

laser trackers, laser radars and total stations

(Cuypers et al 2009)

Finally, instead of using two angles and a distance

measurement, it is possible to evaluate the position of a

point in a three-dimensional (3D) space using angular

information from two or more reference points This

approach relies on the well-known triangulation

princi-ple, which uses the known locations of two or more

reference points and the relative measured angles

between the point to be localised and each reference

point (Doganc¸ay 2005) All camera-based systems

(Mikhail et al 2001) as well as the indoor-GPS (iGPS)

(ARC Second 2010) rely on this working principle

In the last decade, many new large-scale

dimen-sional metrology systems have been designed and

proposed (Peggs et al 2009) In particular, some of

them – composed by several distributed components –

arouse a certain interest because of their unique

features of flexibility and scalability (Maisano et al.2009) In general, according to their architecture,dimensional metrology systems can be classified intotwo groups (see Table 1):

Centralised systems A centralised system isessentially a standalone unit, which worksindependently from other external devices toperform dimensional measurements Often thecentralised unit is difficult to move and itsworking volume is inevitably limited by thetechnology used Laser trackers (AutomatedPrecision 2009, Faro Technologies 2009, LeicaGeosystems 2009), laser radars (Leica Geosys-tems 2009), theodolites and optical CMMs(Axios 3D 2009, Metronor 2009, NorthernDigital 2009, Nikon Metrology 2010) representwidely known examples of metrological instru-ments based on a centralised approach Distributed systems These systems generallyconsist of a set of remotely distributed devicesand a central processing unit, which is in charge

of data acquisition and post-processing tion to provide measurement results According

elabora-to the level of interaction of network devices, afurther distinction can be made between thefollowing categories:

Semi-distributed systems: The distributedapproach is limited to the spatial location ofthe devices, which are just remote sensingunits, providing reference points in the 3Dspace They are unable to communicate withone another and to adaptively reconfiguretheir sensing task The iGPS (Nikon Metrol-ogy 2010) and the V-Star system (GeodeticServices 2010) are representative examples ofsuch a category;

Fully distributed systems: Similar to the distributed ones, these systems are charac-terised by a distributed hardware architecture

semi-In this case, however, the remote sensingdevices are intelligent agents, i.e autonomousentities, which cooperate and coordinate theiractivities to achieve the common objective ofperforming the measurement Owing to theirnature, these systems may be organisedaccording to a flat structure (each node islinked to the central unit) or a hierarchicalstructure (clusters of nodes are in charge ofmore powerful cluster nodes) (Cassandrasand Li 2005)

It has to be observed that, up to now, even if manynoteworthy efforts have been directed towards therealisation of distributed systems, a real fully distrib-uted system has still to come

Figure 1 Working principle of the iGPS To obtain accurate angle measurements, the iGPS uses rotating laser beams (ARCSecond, 2010) Knowing the azimuth and elevation angles (j, y) from two or more transmitters, the system univocally determinesthe position of a probe.

2.1 Distributed dimensional metrology systems

2.1.1 State-of-the-art

The history of distributed LSM systems is quite short

and dates back to the last few years (Peggs et al 2009)

Nevertheless, a few systems – relying on different

technologies and working principles – have already

been proposed (Maisano et al 2009)

The iGPS (Nikon Metrology 2010) uses a number

of distributed transmitters equipped with two rotating

laser beams and an infrared (IR) light to determine the

relative angles from the transmitters to the

photo-diodes sensors embedded in a mobile hand-held probe

(ARC Second 2010) (see Figure 1)

According to the known location of the

transmit-ters, which is normally obtained in an initial setup

phase, the position of the mobile probe can be

calculated (Ferri et al 2010) An experimental

char-acterisation of iGPS operational performance and a

comparison to those of a laser tracker has been

reported in Wang et al (2008), referring to a real

working environment Whereas the distributed system

demonstrated high repeatability, measurement

uncer-tainty resulted to be consistently affected by working

volume size and environmental factors

A similar system is the HiBall (3rdTech 2010) It

consists of two key integrated components: a mobile

probe equipped with lenses and photodiodes and a set of

IR Light Emitting Diodes (LEDs), generally mounted

on the ceiling of the working environment The mobile

probe is able to estimate the angular position of the

LEDs with respect to its local reference system The

position of the probe is thus found by triangulation

knowing the locations of the LEDs (Welch et al 2001)

Developments in imaging technology, which

af-forded the community large-area Charge-coupled

Device (CCD) sensors and improvements of targetimage location algorithms, have led to an ever-increasing competitiveness of vision-based metrology

As a matter of fact, different well-settled solutionsbased on photogrammetry are commercially available,providing for accurate, portable and versatile instru-ments for 3D coordinate-measurement (Optitrack

2009, Vicon 2009, Geodetic Services 2010) Thefundamental principle used by photogrammetry istriangulation (Mikhail 2001) By taking photographs

or video images from at least two different positions, it

is possible to reconstruct the spatial location of a pointand, therefore, the geometry or the main features of anobject Notwithstanding their multiple sensor-basedstructure, the existing photogrammetric instrumentsare applied to reduced working volumes and do notexploit the potentialities of a WSN layout

Generally speaking, although the systems so fardescribed present a physically distributed layout, theirdistributed components do not possess any kind of

‘intelligence’, computational capability or tion potentiality

coordina-2.1.2 Appeal and drawbacksThe appeal of distributed systems derives from manyfeatures that make them different from conventionalcentralised systems:

Flexibility As they consist of multiple remotesensors, distributed systems can be easily ar-ranged in the working volume according to theuser needs, the environment geometry and themeasurement task System flexibility can befurther enhanced by implementing pre-proces-sing software tools providing possible

configuration layout, aimed at optimising the

metrological performance and/or the

measure-ment volume (Franceschini et al 2008, Galetto

and Pralio 2010) The sensor placement problem,

intended as the problem of positioning and

orienting a set of sensors in an attempt to cover

a measurement region, has been extensively faced

referring to different sensing technologies and

fields of application (Younis and Akkaya 2008)

Optimal camera placement issues have been

treated in Mason (1997) and Olague and Mohr

(2002) with reference to accurate 3D

measure-ments through photogrammetric systems The

problem of determining sighting and positioning

strategies for fringe projection sensors is tackled

in Weckenmann et al (2008), to provide

assis-tance to the operator dealing with a

multi-sensors system The next best view problem is

faced in Munkelt et al (2009) as the problem of

planning sensor positions for a 3D

reconstruc-tion task using time-of-flight camera data On

the other hand, the configuration design issue has

been approached also in Ray and Mahajan

(2002) and Laguna et al (2009) for positioning

ultrasonic devices for indoor localisation and

navigation of autonomous vehicles Adaptive

sensor placement strategies (Bulusu et al 2001)

could represent a viable way to provide

capabil-ities to add or remove sensing units according to

the user needs, making these systems extremely

flexible as to their implementation for industrial

applications

Redundancy In typical working conditions,

distributed systems are often able to refer to

more distributed components than strictly

needed Depending on the localisation technique

adopted, information redundancy enhances

sys-tem accuracy and gives the syssys-tem the possibility

of implementing real-time verification strategies

(Franceschini et al 2009b)

Reliability Reliability is the ability of a system

to perform and maintain its functions in

routine circumstances, as well as hostile or

unexpected circumstances A survey of fault

tolerance issues in sensor networks is given in

Koushanfar et al (2004), including techniques

for detection and diagnosis Research work is

focused both on sensor deployment strategies

providing fault tolerance against node failures

(Hao et al 2004; Bredin et al 2005; Han et al

2007) and fault-tolerant algorithms for

detect-ing and localisdetect-ing targets in networks with

faulty nodes (Ding et al 2007) Similarly, if

one or more remote devices are not working

properly, distributed metrology systems, as

generally characterised by hardware dancy, can actually use the ‘healthy’ nodes tocompensate the malfunctioning of a part of thenetwork

redun- Scalabilityredun- The major strong point of a tributed system is the capability to easily adapt

dis-to large dimensions and unusual shapes The realworking volume of a distributed metrologysystem is related to the network layout Chan-ging density and/or position of the remotesensing units, the user can size and shape theworking volume, within the network designphase as well as during the experimentalcampaign

Concurrent measurement capability Distributedmetrology systems generally allow the use ofdifferent measurement tools (multiple targetsand/or mobile probes) at the same time Oncethe system infrastructure has been set up, anunlimited number of tools can actually operatewithin the workspace, without any additionalcost per user

Sensor fusion A wide variety of sensor datafusion applications have been proposed inliterature, considering heterogeneous sensors,spatially distributed sensors and asynchro-nously sampled sensors The use of hetero-geneous sensors, i.e sensors measuring differentphysical entities, entails accurate models forcorrelating sensor observations Moreover, thepossibly distributed sensor architecture requiresmodelling how the observations depend onsensor positioning On the other hand, amodelling of the time evolution of the mea-sured parameters is needed whenever asynchro-nous sampling of sensors is performed.Therefore, a modular architecture could givethe opportunity to integrate the metrologicalsystem with other spatially distributed sensors(in order to monitor temperature, humidity,vibrations, light intensity, etc.) (Akyildiz et al.2007) The sensor data fusion, intended as theprocessing and elaboration of data fromheterogeneous sensors, provides capabilities toperform an environmental mapping of theworking volume and to monitor the operatingconditions of the dimensional measuring de-vices Moreover, multi-resolution systems can

be integrated to provide different granularity indata acquisition

Line-of-sight The distributed nature of theconfiguration layout and the sensor redundancybetter face with ‘visibility’ problems, such asshadowing, line-of-sight obstructions and signalreflections

On the other hand, contrary to centralised systems

that are made by a single metrological unit, the

distributed nature of the described systems requires

the coordination and management of multiple stations

The major disadvantages of these systems are as

follows:

Setup To work properly, every distributed

system needs to know several parameters of the

local hardware Some of these parameters may

change either because of environmental factors

(e.g vibrations, temperature or humidity

changes), or due to unpredictable reasons (such

as accidental movements of network nodes)

Errors during the setup phase adversely affect

the accuracy of the measurements

(Mastrogiaco-mo and Maisano 2010) To achieve the optimum

accuracy, each distributed system generally needs

a careful setup phase During this phase, which

can be automated to some extent, the system

calculates information like positions and

orienta-tions of components, local temperatures,

humid-ity, pressure and so on This information is useful

during the measurement

Expertise Distributed metrology systems are

typically less user friendly than centralised

systems They generally need a more experienced

and careful user, especially during the setup

process Because they consist of multiple

sta-tions, particular attention has to be paid to

coordinate the data acquisition from different

sensing devices (e.g sensor device

synchronisation)

Standards While these new systems are

attrac-tive to potential end-users, standards, best

practice methods and independent performance

verification techniques are usually not available

(Peggs et al 2009)

Accuracy The performance of distributed

me-trology systems is strongly related to several

factors that can affect the accuracy of the system

adversely, such as the number of network

devices, the setup parameters and the network

geometry with respect to the spatial distribution

of measurement points

2.2 WSN–based approach

WSNs are typically composed of small and lightweight

devices that can be easily deployed and arranged in a

working environment Furthermore, each device is

generally provided with both communication and

computation capabilities given by the embedded

electronic components (Akyildiz et al 2002) These

features certainly increase the appeal of WSNs and

make them suitable for the design of a fully distributedsystem Recently, the attention has been focused on theintegration of video devices with scalar sensors, tosetup networks of wirelessly interconnected devicesable to fuse multimedia data from heterogeneoussources Comprehensive surveys on wireless multi-media sensor networks and their applications andtestbeds are available in Akyildiz et al (2007, 2008).Besides existing application fields, such as tracking,home automation, environmental monitoring andindustrial process control, the distributed network-based layout has been more recently implemented also

in dimensional metrology applications (e.g NikonMetrology 2010, 3rdTech 2010) Accordingly, mea-surement systems based on spatially distributed sen-sing units demonstrate profitable scalability features

As a matter of fact, their modular architecture makesthem suitable for LSM, overcoming limitations ofexisting digital photogrammetry-based systems Real-time image acquisition of different targets, possiblylocated in different regions of the working volume, isthen possible by spreading around the sensor devices,provided that the acquisition task is synchronised and

a common reference system is given These capabilitiesmake the proposed system a feasible solution fortracking mobile objects, even if characterised by fastdynamics This property is particularly interesting in

an attempt to automate the contact measurementprocedure Most of the commercially available instru-ments provide an accessory hand-held probe fortouching the reference measurement points (Auto-mated Precision 2009, Axios 3D 2009, Leica Geosys-tems 2009, Nikon Metrology 2010), thus involving adirect interaction between the sensor equipment andthe human operator besides a strong dependence onhis/her skills An alternative approach, proposed inFranceschini et al (2009d), relies on autonomousunmanned platforms for carrying the sensor equip-ment and moving the contact probe around theworking volume According to this new perspective,the human role should be scaled down to simply managethe task-related issues, such as type of measurement (e.g.dimensional measurement, geometry reconstruction andsingle point verification) and data acquisition procedures(e.g point sequences, repeated sampling), and remotelymonitor the unmanned platform that should autono-mously perform the measurement This approach clearlyshows the need for a flexible system architecture that isable to provide measurement data for control as well asmetrological issues

As described in section 2, currently availabledimensional metrology systems rely either on distance

or angle measurements Thus, the possible use of aWSN-based system for dimensional metrology applica-tions is certainly bounded by its capabilities of

performing such kind of measurements Nowadays,

there are many approaches to this field, relying on

different technologies and sensors Angular

measure-ments can be achieved, for example, using

acceler-ometers, magnetacceler-ometers, gyroscopes, CCD sensors,

photodiodes or simply measuring the difference in the

received phase of a radio signal at each element of an

antenna array (Kwakkernaat et al 2008) On the other

hand, distance measurements can be obtained, for

instance, evaluating the ToF of a particular signal

(such as an ultrasound signal), the time difference of

arrival of different signals or the received strength of a

radio communication signal (Franceschini et al 2009c)

Whatever the system components and the

localisa-tion algorithms are, a WSN-based metrology system

represents a further step towards hardware and

soft-ware automation in dimensional measurement

applica-tions Owing to its capabilities of sharing the

metrology task, each network device could work

cooperatively with the aim of determining the

geome-trical features of an object In this way, the

measure-ment results to be the synthesis of the information

locally gathered and shared by each network node

Communication links among the network nodes also

provide capabilities to possibly reconfigure their

orientation during the task according to measurement

conditions and procedures, aiming at optimising the

overall system performance

3 System prototype implementation

The MScMS-II – developed at the Industrial Metrology

and Quality Engineering Laboratory of DISPEA,

Politecnico di Torino – is an indoor

coordinate-measuring system based on IR optical technology and

designed for LSM applications As a first prototype, it

implements both distributed and centralised logics Its

architecture consists of three basic units:

a sensor network of optical devices, suitably

distributed within the measurement volume;

a mobile wireless and armless probe, equipped

with two reflective markers, to ‘touch’ the

measurement points;

a central unit, connected via an antenna linked to

the WSN, to acquire and elaborate the data sent

by each network node

The network of spatially distributed optical sensors

is aimed at providing reference points for locating the

portable probe, by establishing visual links with the

markers that are visible in the camera ‘viewing

volumes’ The probing point, i.e the point of the

probe tip contacting the workpiece, is then calculated

according to the reconstructed positions of markers’

centres and the a priori known probe geometry Anearlier prototype of MScMS exploited ultrasonic (US)transceivers to communicate and evaluate mutualdistances between the distributed sensor nodes andthe hand-held probe (Franceschini et al 2009a) Thepoor characteristics of US devices (e.g non punctiformdimensions, speed of sound dependence on operatingtemperature, wave reflection and diffraction) caused alow accuracy in the measurement results (Franceschini

et al 2009a,b) To enhance system performance,current version implements an IR-based optical out-side-in system, estimating the position of passive retro-reflective markers from their projections in differentcamera views (Galetto et al 2009)

The system is characterised by a fast acquisition andmeasurement procedure, being the acquisition taskrelated to camera frame rates, and the measurementtask dependent on the number of points and theprocessor used System portability strongly depends onmodular architecture, which shares sensing and compu-tational capabilities among several remote units ofreduced size and weight The system can then be easilytransportable and objects can be measured in place.Setup takes just a few minutes, as no warm-up times arerequired for the sensing units The camera self-calibra-tion, based on a collection of images of a single reflectivemarker, randomly moved in several unknown positionswithin the working volume, requires a few minutes,including data gathering and elaboration tasks

3.1 Wireless sensor networkCurrently, the distributed network of the MScMS-IIprototype has been set up by using low-cost commer-cial IR cameras, characterised by an interpolatedresolution of 1024 6 768 pixels (native resolution is

128 6 96 pixels), a maximum sample rate of 100 Hzand a field-of-view (FoV) of approximately 458 6 308.Graphical representations of a single camera viewingvolume and the coverage volume of a set of sensors arereported in Figures 2 and 3, respectively Each cameraimplements a real-time multi-object tracking engine,allowing to track up to four IR light sources To workwith passive markers, each camera is coupled with anear-IR light source (Figure 4), consisting of a 160-chip LED array with a peak wavelength of 940 nm and

a viewing half-angle of approximately 808 The overallsensor set (camera and LED array) weights less than

500 g and is about 13 6 13 6 15 cm size Becausemarker dimensions, camera resolution, IR light sourcepower and working volume are strictly related para-meters, the IR sensor sensitivity has been experimen-tally evaluated by testing the visibility distance ofdifferently sized retro-reflective spheres (see Figure 2).Referring to the used IR technology, the system

demonstrated to be able to track a 16-mm diameter

marker in a range between dmin¼ 50 mm and dmax

3500 mm On the other hand, by using a 40-mm

diameter marker the traceability ranges from 300 to

6000 mm Although the upper bound (dmax) of this

range represents a limitation in terms of marker

visibility in the camera image plane, the lower bound

(dmin) represents the distance under which the tracking

engine is unable to correctly find the centre of the point

projection in its view plane

Besides the camera, each network device isprovided with an accelerometer used for diagnosticpurposes (i.e to detect possible movements or vibra-tions changing the calibrated positions)

It has to be noted that part of the data elaborationtask is locally performed by the remote sensing units

In detail, wireless network devices are in charge of thefollowing:

Image processing To save computational abilities and to reduce the radio communicationloads, all images acquired by the wireless devices

cap-Figure 2 Graphical representation of the sensitivity range of the IR camera The vertical and horizontal view angles areindicated as aVand aH, respectively They identify the camera FOV The light grey volume represents the camera viewing volume,within which a reflective marker is visible and traceable

Figure 3 Graphical representation of the working layout

The dark grey regions represent the ideal viewing volume of

each camera The light grey region identifies the coverage

volume, i.e the volume wherein it is possible to reconstruct

the 3D position of a marker It has to be noted that,

according to triangulation principles, the coverage volume

has been referred as the volume of intersection of at least two

are processed onboard Each device implements

a real-time tracking engine, allowing tracking up

the brightest IR sources in the camera field of

view and thus providing to the central unit their

2D position coordinates

Image filtering To prevent noisy measurements

and undesired reflections, the network devices

only track IR sources with brightness larger than

a certain threshold that is empirically established

No filtering on the shape of the light sources is

performed

As a matter of fact, the embedded real-time

tracking and filtering capabilities of the distributed

remote sensing devices save the computational effort

for performing the image analysis and spot coordinates

identification by the central unit

3.2 Measuring probe

The mobile hand-held probe (Figure 5) consists of a

rod, equipped with two reflective markers at the

extremes and a calibrated tip to physically ‘touch’ the

measurement points Reflective markers have been

made by wrapping a retro-reflective silver transfer film

around polystyrene spheres

Referring to Figure 5, spatial coordinates of point

V can be easily determined, knowing the positions of

the two marker centres and the geometry of the probe,

through a linear equation (Franceschini et al 2009a,

Galetto et al 2009) A correction algorithm, taking

into account the probe trajectory when approaching

the measurement point, has been implemented To this

end, the central processing unit stores the time history

of the reconstructed spatial coordinates of the probe

markers The probe trajectory in a predefined time

interval before the measurement is thus reconstructed

and used to correct the coordinates of the probe tip

3.3 Central unitThe central unit consists of a PC equipped with anIntel Quad Core Q9300 (4 6 2.5 GHz) CPU and 4GBDDR 2 RAM The PC is connected to the WSNdevices via a radio link

The central unit is currently used both for dataprocessing and visualisation The centralised unit is incharge for the following tasks:

Synchronisation According to the radio munication link, cameras are sequentiallysampled This introduces a delay between theacquisition time of images from differentcameras On the other hand, the onboardimage processing lightens the communicationload, thus reducing the acquisition delay to aminimum

com- Camera calibration A camera calibrationprocedure has to be carried to provide to thetriangulation algorithm the spatial coordinates

of the reference points, i.e the IR sensordevices The multi-camera calibration problem

is faced by using a fully automatic technique ofself-calibration (Svoboda et al 2005) Thismethod is able to reconstruct camera internalparameters besides its positions and orienta-tions It requires a minimum of three camerasand a calibrated artefact to align and scale thereference system

Localisation According to the 2D coordinates ofthe IR spot(s) provided by each camera, thecentral unit reconstructs the 3D position of anymarker by applying triangulation algorithms(Hartley and Zisserman 2004)

Pre-processing The central unit is able to runpre-processing software tools to provide anoptimal WSN configuration, taking into accountthe metrology task, the geometry and shape ofthe measurand and the working environmentlayout

Diagnostics The central unit is responsible forrunning diagnostic algorithms, reporting possiblemalfunctions of the sensor devices

These functions have been implemented into adhoc developed software, providing a comprehensiveand user-friendly graphical interface for managingthe measurement tasks (Figure 6) Network designand calibration, marker localisation as well asdimensional measurement tasks are managed throughdifferent applets As a matter of fact, spatialcoordinates of single points, dimensional data aswell as geometric features of the measured object areprovided for both on-line and off-line analysis

Figure 5 Mobile hand-held measuring probe

3.4 Experimental tests for preliminary metrological

characterisation

Preliminary experimental tests have been carried out to

evaluate prototype performance (in terms of

measure-ment accuracy, repeatability and reproducibility) as

well as system potentialities

The data herein discussed refers to a network

layout consisting of six wireless IR cameras, arranged

in a 5.0 6 6.0 6 3.0 m working environment

accord-ing to a grid-based configuration Cameras were

oriented towards the centre of the laboratory, in order

to ensure better volume coverage and measurements

redundancy Figure 3 shows a virtual reconstruction of

the experimental setup

The black wireframe represents the camera

‘field-of-sensing’, whereas the light grey wireframe represents

the working volume (interpreted as the volume of

intersection of at least two ‘field-of-sensing’) It has to

be noted that the actual working volume was about

2.0 6 2.0 6 2.0 m in width

A first evaluation of the measurement accuracy of

point coordinates, intended as the ‘closeness of

agreement between a measured quantity value and a

true quantity value of a measurand’ (VIM 2008), has

been carried out using a 3D aluminium alloy calibrated

artefact (see Figure 7) To have a set of reference points

with known nominal positions, 22 points on the

artefact have been calibrated (at nominal temperature

T¼ 218C and relative humidity RH ¼ 27%) using a

coordinate-measuring machine (DEA IOTA 0101)

The reference points have been measured bythe MScMS-II system prototype, by moving theartefact in five different positions uniformly distrib-uted within the working volume Figure 8 shows thehistogram of the distances between measured andnominal positions

It is noteworthy how the 50% of the measuredpoints is within a distance of 1.87 mm from thenominal position, while the 94.2% of results is lessthan 5-mm far from the nominal position (Galetto

Figure 6 A screenshot of the graphical user interface for managing the dimensional measurement task (example of a planereconstruction task)

Figure 7 The reference artefact used for the accuracy test

et al 2009) At worst, the maximum measured distance

is below 6.5 mm By considering several issues (e.g

geometric distortion of the reconstructed working

volume and measurement process) whose effects on

measurements strongly depend on the location within

the working volume, the severe experimental testing

procedure consistently affect the extent of

measure-ment errors as well as their high variability

In a second test, measurement repeatability of

point coordinates, intended as ‘closeness of the

agreement between the results of successive

measure-ments of the same measurand carried out under the

same conditions of measurement’ (VIM 2008), has

been tested on five different points, uniformly

dis-tributed within the working volume, by repositioning

the probe in the same positions for k ¼ 30 times It is

noteworthy that repeatability characteristics are

re-lated to the sensor device performance as well as to the

operator skills Human skills actually represent an

external factor related to capabilities in holding the

probe in a fixed position The sample standard

deviation of repeatability tests was found to be smaller

than 1.25 mm (Galetto et al 2009)

Finally, measurement reproducibility of point

coordinates, intended as ‘closeness of the agreement

between the results of successive measurements of the

same measurand carried out under changed conditions

of measurement’ (VIM 2008), has been tested with

reference to five points, repeating the measurement

k¼ 30 times with different mobile probe orientations

Reproducibility tests stress the importance on

mea-surement quality of network and probe relative

position and orientation A sample standard deviation

smaller than 3.45 mm has been obtained (Galetto et al.2009)

According to the results emerging from these tests,MScMS-II prototype do not appear to be verycompetitive if compared with commercial systemssuch as CMMs, laser trackers, iGPS With thosetechnologies, in the same working volume, accuracydeviation may range from few micrometers up to 1 mm

at worst, depending on the system and the workingconditions (Franceschini et al 2009a, Maisano et al.2009)

However, these results become particularly esting if cost and potentiality of MScMS-II areconsidered While ensuring scalability and flexibilitythat existing commercial systems cannot guarantee, theprototype still has significant room for enhancementmainly related to the improvement of the employedtechnology Because the state-of-the-art of IR camerasactually provides a wide choice of resolution (from lessthan 1 megapixel up to 16 megapixels), current CCDsensors (128 6 96 pixels of native resolution) could beeasily replaced with higher performance ones Com-mercially available solutions generally enable

inter-Figure 8 Accuracy in distance measurement Histogram of

the distances between measured and nominal positions

Figure 9 (a) Toy car model (b) Measured points andreconstructed shape

intelligent features such as on-board 2D image analysis

and processing, making the computational workload

almost independent of the IR sensor resolution

Nonetheless, a trade-off between the target system

performance and the economic impact of the entire

system has to be found

Furthermore, because of its ease-of-use and fast

data acquisition characteristics, the MScMS-II can be

applied also for geometry reconstruction and reverse

engineering tasks by untrained operators As an

example, the chassis of a toy car model has been

geometrically reconstructed through a triangle-based

linear interpolation relying on 690 measured points

gathered by an unskilled operator in about 1 hr (see

Figure 9)

4 Conclusions

This article discusses the concept of distributed systems

for large-scale dimensional metrology tasks These

systems – due to their nature – are more flexible and

suitable than centralised systems Furthermore, if they

are composed by ‘intelligent’ sensing units, they can

implement network logics (such as auto-diagnostics,

compensation, correction, substitution, etc.) that can

improve the overall performance of the system itself

The proposed approach, based on WSNs,

demon-strates to have the potentialities of being fully

distributed and combining flexibility and network

logics

A prototype implementation, based on a WSN of

IR cameras, has been used to test system feasibility and

demonstrated satisfactory metrological performance

and appealing flexibility, and scalability properties

The prototype represents a step towards fully

dis-tributed systems, because it implements a fully

distributed approach for measurement data acquisition

and both centralised and distributed logics for data

elaboration and sensor management

Future research efforts will go in the direction of a

self-coordination of the remote sensor devices as to

diagnostics and compensation Although accuracy

resulted to be relatively good in view of the technology

used, further studies are intended to increase the

resolution of the optical devices, in order to enhance

the accuracy and working volume coverage

References

Akyildiz, I.F., Melodia, T., and Chowdhury, K.R., 2007 A

survey on wireless multimedia sensor networks

Compu-ter Networks, 51 (4), 921–960

Akyildiz, I.F., Melodia, T., and Chowdhury, K.R., 2008

Wireless multimedia sensor networks: applications and

testbeds Proceedings of the IEEE, 96 (10), 1588–1605

Akyildiz, I.F., et al., 2002 Wireless sensor networks: a

survey Computer Networks, 38 (4), 393–422

ARC Second, 2010 Product literature [online] Availablefrom: http://arcsecond.com [Accessed 24 October 2009].Automated Precision, 2009 Automated Precision Inc.website [online] Available from: http://www.apisensor.-com/tracker3.html [Accessed 4 November 2009].Axios 3D, 2009 Axios 3D Services GmbH website [online].Available from: http://www.axios3d.de/ [Accessed 4November 2009]

Bredin, J.L., et al., 2005 Deploying sensor networks withguaranteed capacity and fault tolerance In: Proceedings

of the 6th ACM international symposium on Mobile adhoc networking and computing, New York: ACM, 309–319

Bulusu, N., Heidemann, J., and Estrin, D., 2001 Adaptivebeacon placement Proceedings of the 21st internationalconference on distributed computing systems, Los Alami-tos, CA: IEEE Computer Society, 489–498

Cassandras, C.G and Li, W., 2005 Sensor networks andcooperative control European Journal of Control, 11 (4–5), 436–463

Cuypers, W., et al., 2009 Optical measurement techniquesfor mobile and large-scale dimensional metrology Opticsand Lasers in Engineering, 47 (3–4), 292–300

Ding, M., et al., 2007 Fault-tolerant target localization insensor networks EURASIP Journal on Wireless Com-munications and Networking, 2007 (1), 1–9

Doganc¸ay, K., 2005 Bearings-only target localization usingtotal least squares Signal Processing, 85 (9), 1695–1710.Faro Technologies, 2009 Faro Technologies Companywebsite [online] Available from: http://www.faro.com[Accessed 16 April 2009]

Ferri, C., Mastrogiacomo, L., and Faraway, J., 2010 Sources

of variability in the set-up of an indoor GPS tional Journal of Computer Integrated Manufacturing, 23(6), 487–499

Interna-Franceschini, F., Mastrogiacomo, L., and Pralio, B., 2009d

An unmanned aerial vehicle-based system for large scalemetrology applications International Journal of Produc-tion Research, 48 (13), 3867–3888

Franceschini, F., et al., 2009a Mobile spatial coordinatemeasuring system (MScMS) – introduction to the system.International Journal of Production Research, 47 (14),3867–3889

Franceschini, F., et al., 2009b On-line diagnostics in themobile spatial coordinate measuring system (MScMS).Precision Engineering, 33, 408–417

Franceschini, F., et al., 2009c A review of localizationalgorithms for distributed wireless sensor networks inmanufacturing International Journal of Computer Inte-grated Manufacturing, 22 (7), 698–716

Franceschini, F., et al., 2008 The problem of distributedwireless sensors positioning in the mobile spatialcoordinate measuring system (MScMS) In: Ninethbiennial ASME conference on engineering systems designand analysis ESDA08,7–9 July 2008 Haifa, Israel NewYork: ASME, 317–327

Galetto, M and Pralio, B., 2010 Optimal sensor positioningfor large scale metrology applications Precision En-gineering, 34 (3), 563–577

Galetto, M., Mastrogiacomo, L., and Pralio, B., 2009 Aninnovative indoor coordinate measuring system forlarge-scale metrology based on a distributed IR sensornetwork In: Proceedings of ASME 2009 internationalmanufacturing science and engineering conference

ASME, CD-ROM

Ganci, G and Handley, H., 1998 Automation in

video-grammetry International Archives of Photogrammetry

and Remote Sensing, 32 (5), 53–58

Geodetic Services, 2010 Geodetic Services Inc website

[online] Available from: http://www.geodetic.com

[Ac-cessed 8 February 2010]

Han, X., et al., 2007 Fault-tolerant relay nodes placement in

heterogeneous wireless sensor networks In: Proceedings

of the 26th IEEE/ACM joint conference on computers and

communications (INFOCOM’07), 6–12 May,

Ancho-rage, AK, Piscataway, NJ: IEEE, 1667–1675

Hao, B., Tang, H., and Xue, G., 2004 Fault-tolerant relay

node placement in wireless sensor networks: formulation

and approximation In: Proceedings of the Workshop on

High Performance Switching and Routing (HPSR’04),

19–21 April Phoenix, AZ, Piscataway, NJ: IEEE, 246–

250

Hartley, R.I and Zisserman, A., 2004 Multiple view geometry

in computer vision Cambridge: Cambridge University

Press

Hofmann-Wellenhof, B., Lichtenegger, H., and Collins, J.,

1997 GPS: theory and practice Wien: Springer-Verlag

Koushanfar, F., Potkonjak, M., and

Sangiovanni-Vincentel-li, A., 2004 Fault tolerance in wireless sensor networks

in Handbook of Sensor Networks.Boca Raton, FL: CRC

Press, 812–829

Kwakkernaat, M.R.J.A.E., et al., 2008 High-resolution

angle-of-arrival measurements on

physically-nonstation-ary mobile radio channels IEEE Transactions on

Antennas and Propagation, 56 (8), 2720–2729

Leica Geosystems, 2009 Leica Geosystems Company

web-site [online] Available from:

http://www.leica-geosys-tems.com [Accessed 4 November 2009]

Laguna, M., et al., 2009 Diversified local search for the

optimal layout of beaconsin an indoor positioning

system IIE Transactions, 41, 247–259

Maisano, D., et al., 2009 Comparison of two distributed

large volume measurement systems: MScMS and iGPS

Proceedings of the Institution of Mechanical Engineers

Part B: Journal of Engineering Manufacture, 223 (5), 511–

521

Mason, S., et al., 1997 Heuristic reasoning strategy for

automated sensor placement Photogrammetric

Engineer-ing and Remote SensEngineer-ing, 63, 1093–1102

Mastrogiacomo, L and Maisano, D., 2010 Network

localization procedures for experimental evaluation of

mobile spatial coordinate measuring system (MScMS)

The International Journal of Advanced Manufacturing

Technology, 48 (9), 859–870

Metronor, 2009 Metronor Corporate website [online]

Available from: http://www.metronor.com [Accessed 10

June 2009]

Mikhail, E.M., Bethel, J.S., and McGlone, J.C., 2001

Introduction to modern photogrammetry New York,

NY: Wiley

Munkelt, C., et al., 2009 View planning for 3D

reconstruc-tion using time-of-flight camera data DAGM 2009,

LNCS 5748, 352–361

Nikon Metrology, 2010 Nikon Metrology website [online].Available from: http://www.nikonmetrology.com [Ac-cessed 26 February 2010]

Northern Digital, 2009 Northern Digital Inc website[online] Available from: http://www.ndigital.com/in-dustrial/products-pcmm.php [Accessed 4 November2009]

Olague, G and Mohr, R., 2002 Optimal camera placementfor accurate reconstruction Pattern Recognition, 35,927–944

Optitrack, 2009 NaturalPoint Company website [online].Available from: http://www.naturalpoint.com/optitrack[Accessed 3 December 2009]

Peggs, G.N., et al., 2009 Recent developments in large-scaledimensional metrology Proceedings of the Institution ofMechanical Engineers, Part B, Journal of EngineeringManufacture, 223 (6), 571–595

Puttock, M.J., 1978 Large-scale metrology Annals of CIRP,

21 (1), 351–356

Ray, P.K and Mahajan, A., 2002 A genetic based approach to calculate the optimal configuration ofultrasonic sensors in a 3D position estimation system.Robotics and Autonomous Systems, 41, 161–177.Svoboda, T., Martinec, D., and Pajdla, T., 2005 Aconvenient multi-camera self-calibration for virtual en-vironments PRESENCE: Teleoperators and VirtualEnvironments, 14 (4), 407–422

algorithm-3rdTech, 2010 3rdTech Inc website [online] Available from:http://www.3rdtech.com [Accessed 2 February 2010].Vicon, 2009 Vicon Inc website [online] Available from:http://www.vicon.com [Accessed 3 December 2009].VIM, 2008 International vocabulary of basic and generalterms in metrology Geneva, Switzerland: InternationalOrganization for Standardization

Wang, Z., et al., 2008 Experimental deployment of theindoor GPS large volume metrology system in a largescale production facility In: Proceedings of the thirdinternational conference on manufacturing engineering(ICMEN), 1–3 October 2008 Chalkidiki, Greece, 827–836

Weckenmann, A., Hartmann, W., and Weickmann, J., 2008.Model and simulation of fringe projection measurements

as part of an assistance system for multi-componentfringe projection sensors Proceedings of SPIE, 7102,71020N-1–71020N12

Welch, G., et al., 2001 High-performance wide-area opticaltracking the HiBall tracking system Presence: Teleo-perators and Virtual Environments, 10 (1), 1–21

Younis, M and Akkaya, K., 2008 Strategies and techniquesfor node placement in wireless sensor networks: a survey

Ad Hoc Networks, 6, 621–655

INFELT STEP: An integrated and interoperable platform for collaborative CAD/CAPP/CAM/

CNC machining systems based on STEP standardOmid F Valilai* and Mahmoud Houshmand

Department of Industrial Engineering, Sharif University of Technology, Tehran, Iran(Received 10 February 2010; final version received 22 September 2010)Integrated product development is comprised of CAD, process planning and CNC code generation based on anintegrated data structure To make various CAD/CAM software solutions with different internal product datastructures interoperable, it is necessary to realise reliable and robust information exchange capability in amanufacturing environment To enable interoperability amongst integrated product development processes,management of collaboration of diverse CAD/CAM software packages is of outmost importance In this article, thefundamental requirements for achieving interoperability and collaboration management in an integrated, enterprise-wide, product development process environment are discussed Based on these requirements, the present prominentintegrated, interoperable and collaborative CAD/CAM information system platforms are comparatively reviewed.There are two main paradigms for development of such platforms – viz neutral-file based and procedure based Themerits and limitations of each paradigm are discussed To overcome the limitations of the current present platforms, athree-layered integrated and interoperable platform, named INFELT STEP, for collaborative and interoperableproduct design/development using different and diverse CAD/CAPP/CAM/CNC software is proposed This platformconjugates the capabilities of both main paradigms to eliminate their limitations The layered structure of INFELTSTEP caters for the requirements of an integrated, interoperable and collaborative computer-based manufacturing andsupports the entire range of software packages in the CAD/CAPP/CAM product development chain In this platform,each software package can send product data based on its own internal data structures to other packages that are linked

to and communicate with this platform Different layers of INFELT STEP convert software packages’ processed data

to its structured data models, which are based on the STEP standard-and then store them in its database Conversely,INFELT STEP layers can retrieve structured data stored in its database and convert them to the format acceptable by aspecific CAD/CAM/CAPP software package that is cooperating in a production workflow This proposed platformmanages the collaboration of CAD/CAM software packages, maintains the integration of CAD/CAM/CNCoperations based on STEP data models, and also enables interoperability of such packages with different local datastructures The capabilities of INFELT STEP are demonstrated by using it in a prototype implementation

Keywords: CAD/CAM integration; ISO 10303 (STEP); manufacturing interoperability; manufacturingcollaboration; distributed product design and production

Introduction

Software frameworks and programs that facilitate

distributed product design and manufacturing are

becoming more and more important in product

development processes (Fenves et al 2003, Panchal

et al 2007) Different solutions and software have been

developed based on these frameworks including

different CAD/CAPP/CAM software packages

Tech-nologies developed for CAD/CAPP/CAM software

packages and CNC post processors are customised

within each of their own application domains named as

automation islands (Xu et al 2005, Gielingh 2008, Xu

2009) So, the application of these software tools in

different enterprises will cause trouble when it is

necessary to exchange product data among engineers

and designers, who are geographically spread and have

different goals, knowledge, experiences, tools andresources, to support collaborative product develop-ment within an integrated product data structure (Pengand Trappery 1998, Kemmerer 1999, Pratt et al 2005,Arthaya and Martawirya 2008) Requirements forcooperative product development can be classified inthree groups:

(1) Managing collaboration of different CAx ware packages in the product developmentprocesses

soft-(2) Enabling interoperability among different CAxsoftware application tools for product dataexchange

(3) Integrating product data spread over multipleenterprise-wide product development processes

*Corresponding author Email: Omidf@ie.sharif.edu

Vol 23, No 12, December 2010, 1095–1117

ISSN 0951-192X print/ISSN 1362-3052 online

DOI: 10.1080/0951192X.2010.527373

http://www.informaworld.com

while they are being updated by different CAx

software packages

To fulfil these requirements, researchers have

proposed different CAx platforms and information

sys-tems These platforms and information systems have

tried to fulfil the above requirements but fall short to

provide a comprehensive all-in-one solution Qin and

Wright (2004) introduced an information system withthe ability to manage the CAx software packagescollaboration and integrated data structure, but it wasnot an interoperable platform and lacked the ability toenable product data exchange among different CAxsoftware packages Xu et al (2005), Xu (2009) andNassehi et al (2006) suggested information systems withthe ability for product data integration based on the

Table 1 Review of important CAD/CAPP/CAM/CNC information system platforms

Platform

Interoperableplatformfor CAD

Interoperableplatform forCAPP/CAM

Management

of CAxand CNCcollaboration

SupportingdistributedCAx and CNCmachiningenterprises

Capability ofCAx and CNCmachiningintegrationbased

on STEP

Ability tosupport newCAxapplicationsystemsMulti-Agent System for

Computer Aided Process

STEP based Engineering Data

Management (EDM) system

(Peng and Trappery 1998)

(Lee and Jeong 2006)

Intelligent agent System

architecture (Marchetta and

CAE system architecture for

support multiple viewpoints

Intelligent CAD/CAM system

VIVACE – EDM architecture

(Nguyen Van et al 2007)

(Newman and Nassehi 2007)