You are on page 1of 6

Available online at www.sciencedirect.

com

ScienceDirect
Procedia CIRP 57 (2016) 398 – 403

49th CIRP Conference on Manufacturing Systems (CIRP-CMS 2016)

A Model based Visualization Framework for Cross Discipline


Collaboration in Industry 4.0 Scenarios
Johannes Hertera*, Jivka Ovtcharovab
a
UNITY AG, Wankelstrasse 3, 70563 Stuttgart, Germany
b
Karlsruhe Institute of Technology, Institute for Information Management in Engineering, Zirkel 2, Bldg. 2020, 76131 Karlsruhe, Germany
* Corresponding author. Tel.: "+49 162 88 255 36" ; fax: +49 711 686 890 59. E-mail address: johannes.herter@unity.de

Abstract

The Industry 4.0 initiative fosters close collaboration between various disciplines in product development. Collaborative aspects are addressed
by several research activities, namely Systems Engineering and more precisely by Model Based Systems Engineering. This work contributes a
new method for interdisciplinary communication by an integrated visualization of conceptual models. The method describes an environment for
direct communication between experts in related disciplines. Common application scenarios are business process management workshops in
which process experts have to achieve commitment with experts from other disciplines like computer science, logistics or engineering. Especially
in Industry 4.0 scenarios, mutual understanding between experts from several domains is crucial for a successful process setup, planning and
preparation of production scenarios. The method comprises a theoretical framework for establishing inter-model relationships and their
visualization. For establishing links between models, an ontology based approach is applied. The Bunge-Wand-Weber Ontology for information
systems provides a generic and yet applicable basis for establishing consistent inter-model relationships. For supporting communication processes
the integrated models should be displayed. To provide an understandable visualization a three-dimensional environment is proposed, in which
the linked models are arranged and displayed. The scientific basis is given by applying the concept of visual variables described in the scientific
discipline of information visualization. The depth of space extends the set of visual variables. Applied to the semantic concept of integrated
models a discriminable visualization of integrated models is achieved.
© 2016 Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
© 2015 The Authors. Published by Elsevier B.V.
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of Scientific committee of the 49th CIRP Conference on Manufacturing Systems (CIRP-CMS 2016).
Peer-review under responsibility of the scientific committee of the 49th CIRP Conference on Manufacturing Systems
Keywords: Industry 4.0; Collaboration; Information Visualization

1. Introduction A common practice for problem solving in technical


domains is to follow a model-based approach. Experts apply
Recent development in industrial production scenarios are domain specific conceptual models to formalize their
dominated by a rising complexity of product technologies and understanding. Thereby the complexity of real world scenarios
manufacturing techniques. Principals of the Industry 4.0 is reduced, as models comprise of a selection of relevant
initiative postulate a close integration of physical components elements. (De-facto) Standardized conceptual model grammars
and IT systems for developing and applying cyber-physical prevail most technical domains. Model grammars describe
technical systems in product manufacturing scenarios. The constructs and relationships, which are relevant for problem
technical components of cyber-physical systems are solving in their native domain. Some grammars define a visual
increasingly sophisticated. This demands deep knowledge of notation for a graphic representation of the semantic model.
domain experts who participate in the product development Established modelling grammars like BPMN (Business
process. The results of participating domains have to be Process Modelling and Notation), UML (Unified Modelling
integrated to establish a sound and usable technical product. Language) or SysML (Systems Modelling Language) can
achieve the status of a common language or shared mental

2212-8271 © 2016 Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the scientific committee of the 49th CIRP Conference on Manufacturing Systems
doi:10.1016/j.procir.2016.11.069
Johannes Herter and Jivka Ovtcharova / Procedia CIRP 57 (2016) 398 – 403 399

model as they are widely known and applied. This is a great perspectives in product engineering, I4.0 scenarios define the
benefit for collaboration within the domain, however there is organizational scope vaster. I4.0 scenarios furthermore
little benefit for inter-domain collaboration. consider value networks as the organizational framework for
Specific conceptual models are not necessarily known developing smart products on the basis of integrated production
outside their native domain and they cannot be used as a technologies and methods. This poses new demands for
common language. However, interdisciplinary communication collaboration between organizations, hierarchies and technical
requires users to achieve common ground for mutual domains. In 2014 a reference architecture titled RAMI was
understanding [1]. A method to support knowledge exchange introduced [5] to describe an architecture as basis for deriving
between domains based on visually linked conceptual models use-cases and further discussions. It describes multiple layers
could enhance mutual understanding. Still, experts can stick to and domains which collaborate closely for successful product
their familiar mental model and notation. Statements like: “… development.
that’s what I meant to say” should be made earlier with suitable Most of the addressed domains apply conceptual models for
communication support based on a good visualization. performing their daily tasks, throughout the product lifecycle
as well as in organizational hierarchies. This shows a
1.1. Objectives significant demand for a method to support communication and
collaboration between stakeholders.
This work contributes to the body of knowledge a method
to support model-based communication in interdisciplinary 2. Conceptual model integration
collaboration scenarios. Domain experts should be able to use
their native conceptual models and notation. The relationships Conceptual models are common structures for capturing and
between conceptual models should be transparent and communicating knowledge of a domain. Definitions describe
intuitively understandable. This should be achieved by visual some fundamental characteristics of conceptual models which
means. The method should cover the following: are comprehensively noted in General Model Theory [6]:
x An approach for selecting and linking elements of several x Mapping: A conceptual model is a reproduction of an
conceptual models based on semantics. original. The original might be a real world scenario or
x A method to provide a sound and intuitive visual another model [6].
representation for inter-model relationships. x Reduction: The conceptual model consists of a subset of
Both aspects should be based on scientific theories. elements contained in the original. Only relevant entities
The work introduced here is developed in scope of a are adopted in the model [6].
doctoral thesis, which is published in 2016 at Karlsruhe x Pragmatism: The selection of artefacts in the model is
Institute of Technology. This paper gives an overview of driven by the purpose of the model and the information
fundamentals and implementation of a conceptual framework demand of the modeler. The pragmatism determines the
published earlier [2,3]. The concept for a visual representation level of reduction and the structure of the model [6]
of inter-model relationships in the third dimension was The reduction and pragmatism indicate the difficulties of
addressed in [4]. domain-specific conceptual models in cross-discipline
This paper focusses on the ontology-based approach for collaboration scenarios: the understanding and information
model integration and gives insights about the implementation demand inside a domain determines the pragmatism.
and evaluation of the conceptual framework. Domain-specific conceptual models most likely define
varying entities as the models are developed with specific
1.2. Structure and Scope pragmatisms.
This results in heterogeneous conceptual models. Efforts for
This paper is structured as follows: Initially the issue of information integration become necessary to support
cross-domain collaboration is put in the context of Industry 4.0. information exchange between models.
This is followed by introducing the fundamentals of semantic
model integration. On this basis the concept for establishing 2.1. Integrating conceptual models for a common
inter-model relationships is described. The fourth section visualization
introduces the essentials of information visualization with
focusing on concepts for graphic representation of conceptual The challenge to resolve heterogeneity between conceptual
models. Both approaches for model integration and models seems to be a general issue in various domains. The
visualization are integrated in section five which briefly heterogeneity is caused by different data formats and
introduces the prototypical implementation. The final section structures, as well as varying semantics of elements in models
gives an overview about the results of empirical and theoretical [7]. Efforts were conducted for integrating conceptual models.
evaluation. The conclusion then focusses on future application Two general approaches seem to give promising results:
scenarios in the context of Industry 4.0. x Applying singular models which cover concepts from
several domains, e.g. MBSE (system level) [8], STEP [9],
1.3. Application Scenario internal models of ERP Systems

The Industry 4.0 (I4.0) initiative fosters the interconnection


of domains and technologies. Despite of current and past
400 Johannes Herter and Jivka Ovtcharova / Procedia CIRP 57 (2016) 398 – 403

x Mechanisms for model integration by mapping and


translation ([10,11]).

2.2. Semantic integration

A common technique to resolving semantic heterogeneity is


applying ontologies for formalizing knowledge in conceptual
models [10]. The ontology is a method for semantic based
translation between conceptual models. The ontology
comprises of concepts, attributes and relationships, which
express the general architecture of a conceptual model. The
ontology in this case has the role of a meta-model, a structure
which describes the semantics of entities inside a conceptual
Fig. 1 BWW-Properties and model attributes
model [12].
An established ontology for analyzing information systems
ontology is applied to an information-system comprising of
is the Bunge Wand Weber (BWW) ontology [13]. Although it
several related conceptual models.
has also been criticized [14], it is widely applied to verify the
The ontology based approach for establishing inter-model
quality of conceptual models especially in the domain of
relationships is quite generic, however it provides a profound
industrial engineering [12,15–17].
scientific basis for explaining the selection of entities which are
The formal BWW-ontology defines generic constructs and
linkable.
connections of elements, which are commonly used in
The occurrence of inter-model relationships based on shared
technical information systems. They are not bound to any
BWW-Properties requires annotation in the visualization as
specific domain. An overview including a brief explanation of
they are not an integral part of any model. Therefore a type is
the core concepts is summarized by Green and Rosemann [18].
assigned to the shared property which is displayed to users (see
The application scenario to identify links between domain
section 6.2.).
specific conceptual models uses the following ontological
constructs of the BWW Ontology:
3. Information visualization
x Things: Any physical or identifiable element in an
information system
The Information Visualization discipline conducts methods
x Property: A characteristic of a thing, which has not
and techniques for graphic representations of abstract data [19].
necessarily to be part of a model.
An established theory in this field is the visual variable theory.
x Attribute: A Property, which is captured in a model
Originally described by Bertin in context of cartography [20],
x Transformation: A change of status of a thing, this element
this theory defines a set of visual variables, which are used to
is applied for describing processes
represent a conceptual model graphically. The set of visual
variables is comprises of constructs, which can be used to
2.3. Approach for establishing inter-model relationships
picture semantic constructs on a two-dimensional plane of
projection, e.g. shape, orientation, size or color.
The BWW Ontology is rather generic, yet the constructs and
A principle for achieving high quality visualization is to
definitions provide a theoretical framework for selecting and
create a distinct mapping between semantic concepts and visual
establishing relationships between concepts in different
variable. This principle is referred to as semiotic clarity [21]. If
models.
not fulfilled, a semantic construct is represented by several
Key element are the different characteristics of BWW-
visual variables or in opposite, a graphic symbol represents
Properties and BWW-Attributes. BWW-Properties are related
several semantic constructs [21]. As a result users might get
to the real world. They are not necessarily captured in any
confused about the graphic representation. The requirement of
model implementing the ontology. As long as the property can
a distinct mapping of semantic construct and graphic element
be identified in the original, the definition of the property is
demands introducing an unassigned visual variable for the
fulfilled. Properties can depend on the relationship between
semantic construct of inter-model relationships, respectively
things. In this case, they express shared characteristics. BWW-
the common visualization of several models within one
Properties are not necessarily represented in the conceptual
environment.
models, which relate to the original scenario and are formalized
in the models.
4. Visualization of connected conceptual models
In consequence inter-model relationships can be defined on
the existence of shared properties. If concepts of the original
The inter-model relationships have fundamentally different
scenery have shared properties, then an inter-model
characteristics than internal model relationships. Internal
relationship can be established. Inter-model relationships can
relationships link entities within a model and therefore they are
be introduced to the scenario of conceptual models, even
an integral component of the model.
though they are not part of any model. The given rules and
Inter-model relationships are established on the basis of the
definitions of BWW-Ontology are not altered. Basically, the
introduced concept of shared BWW-Properties and are not
Johannes Herter and Jivka Ovtcharova / Procedia CIRP 57 (2016) 398 – 403 401

defined in any model-grammar. As a consequence there is no 5. Implementation and empirical insights


predefined notation.
It is necessary to choose a discriminative notation for The prototypical implementation is based on a virtual reality
internal and inter-model relationships to fulfill the demand for framework developed at Karlsruhe Institute of Technology.
a dedicated mapping between semantic construct and visual PolyVR is an extensible visualization framework which
notation. However, a common visual representation for provides a straightforward scripting environment for creating
relationships between entities are lines [19,22]. and displaying 3D scenes in various hardware environments
For fulfilling the design principle to “speak the user’s [25].
language”[23] and to use a unique mapping of visual variable To build up a scene, graphical models are imported via the
to semantic constructs, an additional visual variable is required. .svg (Scalable Vector Graphics) file format. The PolyVR
implementation renders and displays the models. Inter-model
4.1. Depth as visual variable relationships are read from text files and are added dynamically
to the scene. PolyVR uses the texture coordinates of the
To achieve a distinct graphic representation for inter-model graphical symbols in .svg files to identify the origin and target
relationships we introduce “depth of space” as a visual variable of the inter-model relationships. As most authoring systems for
to a multi-model visualization. The depth of space extends the graphic modelling support the .svg file format and the layout of
classic two-dimensional set originally proposed in [20]. This exported models should be familiar to users. All examples
results in distinct means of design for the special characteristics shown in this paper were drawn in Microsoft Visio and
of a multi-model visualization [4]. Models are arranged in the exported to .svg via internal functionality.
visualization space, which makes them easily distinguishable.
5.1. Model visualization in a common space

For evaluation purposes, model scenes with two and three


models have been implemented. Both conceptual framework
and the current implementation allow arbitrary placement of
models. The parallel and triangular placements were chosen as
the angle between inter-model relationships and model planes
is maximized. This makes the introduced depth component as
a visual language particularly visible.
Depending on the user’s point of view, the front model

Fig. 2 Spatial arranged models

Users can navigate in the spatial scene and focus on


individual models by zooming and altering their point of view.
In collaborative scenarios users do not have to change windows
or switch screens between models which reduces orientation
efforts when focusing on individual models [24].
Inter-model relationships have a depth component, as they
link elements of several models arranged in space. Therefore Fig. 4 Spatial arrangement of integrated models
they are easily distinguishable to internal relationships which
stick to the model plane. covers parts of the model behind (see fig. 2, 3). The spatial
visible effect is nullified. This effect is overcome by a
transparent background of the models. If the models are placed
in appropriate distance and transparency set to an intermediate
level, the text and graphics of the front model are readable, yet
the structure of models behind is visible.
Users can freely navigate within the information space
containing the models to either gain an overview from distance
or analyze sections of singular models in detail.

5.2. Display of Relationships

Inter-model relationships are drawn as straight lines


between the model planes. The line symbol is chosen as it is
the common graphic representation for interconnecting
elements. This supports a guideline for graphic notation
Fig. 3 Inter-Model Relationship
referred to as to “speak the user’s language” [23]. The spatial
402 Johannes Herter and Jivka Ovtcharova / Procedia CIRP 57 (2016) 398 – 403

component as visual variable indicates the inter-model in a model, only the entities are displayed which have their
characteristic of the relationship. The coloring of the lines are origin or target on the clicked artefact. This feature implements
additive means for displaying the inter-model relationships. the widely accepted “mantra” for graphic visualization
“overview first – filter–details on demand” [28].

7. Evaluation

The evaluation of this concept was conducted with


experienced business consultants during a workshop and
several individual discussions.
The evaluation process was conducted as follows: Initially
a real world scenario is described in which users should derive
requirements from a business process. Thereby they should
evaluate, if the derived requirement is realizable in the given
Fig. 5 Display of inter-model relationship type system environment. This task requires switching continuously
between the model diagrams. After performing the evaluation
Users reported during evaluation, that the spatial component of tasks, users were interviewed in a structured query about
the inter-model relationship is the dominant visual clue for personal impression, additional requirements and potential
discriminating the lines within and between the models. application scenarios.
The relationships are added to the integrated models on basis All users mentioned the benefit to have all diagrams in a
of mutual BWW-Properties, which are not integral part of any common environment and not having to switch between
model. Therefore it is necessary to indicate the type of windows or charts. The visible links between the diagrams
relationship, as described in section 3.3. The type of were appreciated for analysis and communication purposes.
relationship is displayed in the space between the models. The Ideas for improvement focused on interaction paradigms and
relationship type is visible, when the user follows an inter- visual cues during navigation. One significant issue identified
model relationship and enters the space between models. is the issue that a model getting out of the field of vision when
entering the space between models. The model formally in the
6. Visual Complexity front is virtually behind the user and therefore not visible
anymore. This impedes detail analysis, as users have to
The visual complexity is a measure to evaluate the remember where they came from when entering the space
understandability of graphic notation [26]. The complexity between models. This could be overcome by implementing a
arises of the number of graphic elements and relationships visual hint (e.g. picture-in-picture) indicating the origin of the
which are displayed together. The visual complexity rises when analyzed inter-model relationships.
several models, including inter-model relationships, are During the discussion a number of potential use-cases were
displayed as shown in fig.5. Depending on the use-case, the been proposed. Examples covered:
display of all relationships can support users in problem solving x Process evolution: A common visualization of process
(overview). Fig. 5 shows an IT-landscape on the left and a models in several stages could make the evolution of the
BPMN [27] Process model on the right hand side. The models transparent and communicable.
overview gives an impression on first sight about the x Mapping of processes to layout: This use-case would
significant model entities for integration aspects by the number shows the location where processes and activities are
of inter-model relationships pointing to them. performed. This use-case is addressed by [3,29] in a
However this representation is not suitable for analysis sophisticated way.
tasks which require a detailed view. To support analysis tasks x Organization- process mapping: International
the implementation provides mechanisms for filtering inter- organizations might have different organizational
model relationships. By clicking on any of the linked entities structures to implement standardized processes. The
displaying of responsibilities in organizations can support
restructuring projects.
x Implementation of a use-case for RAMI architecture for
industry 4.0 applications. The arrangement and
interconnection of models could implement the multi-
dimensional aspects of the reference architecture. The
method introduced here supports arranging and
interconnecting several conceptual models behind each
other. With this feature the multiple layers of a RAMI
implementation can be fully displayed. The visualization
of several layers should support understanding and
communication in a complex Industry 4.0 scenario.
Fig. 6 Filter mechanisms to overcome visual complexity
Johannes Herter and Jivka Ovtcharova / Procedia CIRP 57 (2016) 398 – 403 403

8. Conclusion [9] ISO/IEC, ISO 10303-214 DIS -Part 214 2nd Edition, ISO -
International Organisation for Standardization, 2003.
This paper describes a method to interconnect and visualize [10] V. Bittel, Semantische Informationsintegration − Konzeption eines
conceptual models from different domains which formalize a auf Beschreibungslogiken basierenden Integrationssystems für die
common problem set. This method supports collaborative Produktentwicklung, KIT, 2014.
issues in industrial product development by making the [11] M. Fowler, Patterns of Enterprise Application Architecture, Wiley,
interrelation between the models visible and transparent. In 2003.
industry 4.0 scenarios the demand for collaboration between [12] M. Rosemann, I. Vessey, R. Weber, B. Wyssusek, On the
experts from several domains is expected to increase, indicated applicability of the Bunge-Wand-Weber ontology to enterprise
by the multidimensional RAMI architecture for I4.0 scenarios. systems requirements, in: Australas. Conf. Inf. Syst., 2004.
The method introduced provides a tool to ease collaboration [13] Y. Wand, R. Weber, An Ontological Model of an Information
and accelerate decision making processes based on a spatial System, IEEE Trans. Softw. Eng. 16 (1990) 1282–1292.
visualization of linked conceptual models. [14] B. Wyssusek, Ontological foundations of conceptual modelling
Future work focuses on evaluating the method introduced in reconsidered: a response, Scand. J. Inf. Syst. 18 (2006) 139–152.
industrial application scenarios. [15] J. Becker, P. Bergener, D. Breuker, M. Räckers, Evaluating the
Furthermore we implement a multi-layer model expressiveness of domain specific modeling languages using the
visualization to evaluate several models are arranged behind Bunge-Wand-Weber ontology, Proc. Annu. Hawaii Int. Conf. Syst.
each other for visualizing a multi-layer architecture of models. Sci. (2010).
This contributes to understanding and communicating [16] A. Gehlert, D. Pfeiffer, J. Becker, The BWW-Model as Method
instances of the RAMI reference architecture in Industry 4.0 Engineering Theory, Proc. 13th Am. Conf. Inf. Syst. (AMCIS
scenarios. 2007). (2007) 1–10.
[17] S. Nurcan, C. Salinesi, C. Souveyet, J. Ralyté, Intentional
References Perspectives on Information Systems Engineering, Springer Berlin
Heidelberg, Berlin, Heidelberg, 2010.
[18] P. Green, M. Rosemann, Integrated process modeling: An
[1] H. Clark, S. Brennan, Grounding in communication, in: Perspect. ontological evaluation, Inf. Syst. 25 (2000) 73–87.
Soc. Shar. Cogn., American Psychological Association, [19] C. Ware, Information visualization: perception for design, Morgan
Washington, D.C., USA, 1991: pp. 127–149. Kaufmann, 2004.
[2] D. Eichhorn, J. Herter, A. Oberweis, An Approach for a Domain- [20] J. Bertin, Semiology of Graphics, Esri Press, 2010.
spanning Collaboration Platform for Decision Support Using [21] D. Moody, The “Physics” of Notations: Toward a Scientific Basis
Immersive Visualization Techniques in Product Manufacturing, in: for Constructing Visual Notations in Software Engineering, IEEE
A. Nolte, M. Prilla, S. Lukosch, G. Kolfschoten, T. Herrmann Trans. Softw. Eng. 35 (2009) 756–779.
(Eds.), 1st Int. Work. Collab. Usage Dev. Model. Vis. CollabViz, [22] E.R. Tufte, Envisioning Information, Graph. Press. 79 (1998) 346–
Aarhus, Denmark, 2011. 348.
[3] R. Brown, J. Herter, Virtual World Process Perspective [23] J. Nielsen, Usability Engineering, Morgan Kaufmann, 1994.
Visualization, in: Conf. Information, Process Konwledge Manag., [24] S. Meyers, The Keyhole Problem, Sites J. 20Th Century Contemp.
Valencia, Spain, 2012. French Stud. (2002) 1–25.
[4] J. Herter, R. Brown, J. Ovtcharova, A Visual Language for the [25] V. Häfner, PolyVR - A Virtual Reality Authoring System, in: G.
Collaborative Visualization of Integrated Conceptual Models in Zachmann, J. Perret, A. Amditis (Eds.), Conf. Exhib. Eur. Assoc.
Product Development Scenarios, in: M. Abramovici, R. Stark Virtual Augment. Real., The Eurographics Association, 2014.
(Eds.), Smart Prod. Eng., Springer, 2013: pp. 805–814. [26] S. Harper, E. Michailidou, R. Stevens, Toward a definition of visual
[5] J.H. Henning Kargermann, Wolfgang Wahlster, complexity as an implicit measure of cognitive load, ACM Trans.
Umsetzungsempfehlungen für das Zukunftsprojekt Industrie 4.0, Appl. Percept. 6 (2009) 1–18.
Acatech, 2013. [27] S.A. White, Introduction to BPMN, (2004) 2008–2029.
[6] H. Stachowiak, Allgemeine Modelltheroie, Springer Berlin / [28] B. Shneiderman, The eyes have it: a task by data type taxonomy for
Heidelberg, 1973. information visualizations, Proc. 1996 IEEE Symp. Vis. Lang.
[7] L. Obrst, Ontologies for semantically interoperable systems, Proc. (1996) 336–343.
Twelfth Int. Conf. Inf. Knowl. Manag. - CIKM ’03. (2003) 366. [29] R. Brown, F. Cliquet, Communication of Business Process Models
[8] M. Eigner, R. Daniil, R. Zafirov, Modellbasierte Virtuelle via Virtual Environment Simulations, BPTrends. 12 (2008) 1–7.
Produktentwicklung, 2014.

You might also like