<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD v1.0 20120330//EN" "JATS-archivearticle1.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink">
  <front>
    <journal-meta />
    <article-meta>
      <title-group>
        <article-title>Enhancing Model-Based Engineering of Product Lines by Adding Functional Safety</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <string-name>Stephan Baumgart</string-name>
          <email>stephan.baumgart@volvo.com</email>
          <xref ref-type="aff" rid="aff1">1</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Joakim Froberg</string-name>
          <email>joakim.froberg@mdh.se</email>
          <xref ref-type="aff" rid="aff2">2</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Sasikumar Punnekkat</string-name>
          <email>sasi@goa.bits-pilani.ac.in</email>
          <xref ref-type="aff" rid="aff0">0</xref>
          <xref ref-type="aff" rid="aff2">2</xref>
        </contrib>
        <aff id="aff0">
          <label>0</label>
          <institution>BIT-Pilani KK Birla Goa Campus</institution>
          ,
          <country country="IN">India</country>
        </aff>
        <aff id="aff1">
          <label>1</label>
          <institution>Dept. Change Management and Process Development, Volvo Construction Equipment</institution>
          ,
          <addr-line>Eskilstuna</addr-line>
          ,
          <country country="SE">Sweden</country>
        </aff>
        <aff id="aff2">
          <label>2</label>
          <institution>School of Innovation, Design and Engineering, Malardalen University</institution>
          ,
          <addr-line>Vasteras</addr-line>
          ,
          <country country="SE">Sweden</country>
        </aff>
      </contrib-group>
      <abstract>
        <p>Today's industrial product lines in the automotive and construction equipment domain face the challenge to show functional safety standard compliance and argue for the absence of failures for all derived product variants. The product line approaches are not su cient to support practitioners to trace safety-related characteristics through development. We aim to provide aid in creating a safety case for a certain con guration in a product line such that overall less e ort is necessary for each con guration. In this paper we 1) discuss the impact of functional safety on product line development, 2) describe a model-based approach to capture safety-related characteristics during concept phase for product lines and 3) discuss the usefulness of our proposal.</p>
      </abstract>
      <kwd-group>
        <kwd>Product Line Engineering</kwd>
        <kwd>Functional Safety</kwd>
        <kwd>Model-based</kwd>
        <kwd>Systems Engineering</kwd>
        <kwd>ISO 26262</kwd>
      </kwd-group>
    </article-meta>
  </front>
  <body>
    <sec id="sec-1">
      <title>-</title>
      <p>
        Reuse of already developed components and system parts is commonplace in
industry today and the main goal is to reduce cost and achieve faster time to
market. The industrial product lines we observe in our studies are characterized
by an engineer's mindset and a clone-and-own strategy instead of a managed
and organized reuse in software product line engineering (SPLE) concepts.
Accordingly, the practices around product line engineering have aws in industry
today, i.e. the state of the art practices are not implemented. At the same time
the products developed in the automotive and construction equipment domain
need to ful ll functional safety standards. The functional safety standards like
ISO 262626 [
        <xref ref-type="bibr" rid="ref1">1</xref>
        ] de ne requirements on the development process to avoid
systematic and random failures. Evidence on how potential hazards have been taken
into consideration throughout the development of the product need to be
collected and provided in a safety case. Functional safety compliance is achieved by
applying rigor in the process of developing the system. Copying from other
products or previous product generations would involve that nothing has changed in
the safety argumentation. This is not always the case and instead may lead to
unexplored hazards or violations of safety goals. The exibility in creating
variants can in some cases increase the e ort for assuring compliance. Instead of
just assuring that a component cannot fail dangerously, we now face a situation
where we must assure that no variant can fail dangerously, in any of the
possible con gurations. Many functional safety standards assume a V-model-based
development process without support for product line development. While the
state of the art methods for product line engineering do not encompass methods
and models for achieving functional safety compliance.
      </p>
      <p>
        There is not just one solution on how to set up a product line, instead di erent
product line strategies can be chosen. Jan Bosch describes di erent product
line strategies for software product lines in [
        <xref ref-type="bibr" rid="ref2">2</xref>
        ] and proposes a categorization of
maturity levels for product lines. Applying this categorization on the systems
level implies that each product line category requires a di erent approach to
functional safety. Choosing a product line strategy has an impact on the possible
safety concepts on the one hand and their allocation to technical solutions on the
other hand. Both the distributed development and the diversity of tool chains
hinder the communication about the development in general and in particular
about functional safety. Flawed communication is one reason for potential errors
and systematic failures [
        <xref ref-type="bibr" rid="ref3">3</xref>
        ]. The e ort for achieving functional safety standard
compliance is higher than the actual development e ort for highly safety-critical
single product development already. There is a need to provide guidance and
methods enabling practitioners to manage functional safety in product lines more
e ciently and e ectively.
      </p>
      <p>The contribution of this paper is a model-based approach to manage
functional safety during concept phase in product line development. It is necessary
to start from the systems perspective since functional safety is a property of the
system and we therefore focus on the concept phase described in the functional
safety standard ISO 26262 Part 3.</p>
      <p>The paper is structured as follows. The related work is discussed in 2 and
in section 3 we describe our approach and present a case from the construction
equipment domain. In section 4 we discuss our approach and conclude the paper
in 5.
2</p>
    </sec>
    <sec id="sec-2">
      <title>Background and Related Work</title>
      <p>In our work we aim document functional safety and provide the base to derive
a suitable product line strategy. Typical concepts for documenting functional
safety are Document-based approaches, Architecture Description Languages,
Component-based approaches and Model-based approaches.</p>
      <p>Document-based approaches: It is common in practice to specify the work
products required by the functional safety standards in separate documents. This
is su cient for small and less complex systems with independent safety-critical
functions. Documents can be misinterpreted and misunderstood, which especially
in companies with distributed global development leads to that functional safety
related documents may be interpreted di erently and di erent technical solutions
are developed. Managing the complexity of product lines and functional safety
with a document-based approach is challenging and dependencies may be missed.</p>
      <p>
        Architecture Description Languages The focus of an architecture
description language (ADL) is to describe the architecture of the embedded
system. EAST-ADL [
        <xref ref-type="bibr" rid="ref4">4</xref>
        ] is an ADL, which has been proposed to aid the development
of embedded systems in the automotive domain. It is covering the development
phases from vehicle level onwards where features are documented and variability
of the product line is analyzed and captured. In EAST-ADL2, the extension of
EAST-ADL, an error model and a safety case metamodel are added. The
authors in [
        <xref ref-type="bibr" rid="ref5">5</xref>
        ] de ne safety contracts and propose a set of rules for EAST-ADL2 to
provide automatic proofs if safety goals and safety contracts are violated. Sun
et al. [
        <xref ref-type="bibr" rid="ref6">6</xref>
        ] describe a concept to transform a Product Line Fault Tree (PLFT)
into an AADL (Architecture Analysis and Design Language) model to enable
connecting the hazards to elements in the AADL model. This assumes though
that the product line system is already modeled in order to map the hazards.
Details on how to derive a product line concept under consideration of functional
safety are not yet provided.
      </p>
      <p>
        Component-based Approaches Component-based approaches aim to
describe an embedded system in detail focusing on the software components and
their interaction. The CHESS project [
        <xref ref-type="bibr" rid="ref7">7</xref>
        ] aims to document safety related
information in a component model enabling an automated dependability analysis.
The authors introduce dependability concepts added to the component model.
Product line engineering is not considered in the project and the concept
assumes that all information is available when the component model is used. In
the recent years concepts for mapping hazards to speci c component have been
developed. The authors in [
        <xref ref-type="bibr" rid="ref8">8</xref>
        ] describe a concept for creating component fault
trees (CFT), which aim to map relate parts of the fault tree to the according
components of the design. Gomez et al. [
        <xref ref-type="bibr" rid="ref9">9</xref>
        ] describe the application of the CFT
concept and claim, that e orts for performing a FTA are reduced in the future,
when components are reused. A CFT for a component can rst be derived after
development and therefore bene ts of the approach are rst evident during reuse
of the component.
      </p>
      <p>
        Model-based Approaches Model-based development approaches are
growing importance for developing embedded systems and can be applied for the
systems level (SysML) as well as for the detailed technical descriptions (UML).
Biggs et al. [
        <xref ref-type="bibr" rid="ref10">10</xref>
        ] propose a SysML based approach to capture safety related
information in a model. They assume that all relevant information is available and
describe how to use the SysML diagrams to create a common documentation.
The authors do not describe how to achieve functional safety in product lines
though. Liu et al. [
        <xref ref-type="bibr" rid="ref11">11</xref>
        ] describe a concept to perform a safety analysis for software
product lines and exploring potential hazardous states using UML state chart
diagrams and scenario diagrams. An UML-based approach to model software
product lines is proposed by Gomaa [
        <xref ref-type="bibr" rid="ref12">12</xref>
        ], which is both focusing on the domain
engineering phase, where a common software architecture is derived to support
all relevant product variants and the application engineering phase, where the
common architecture is applied to derive the product variants. Functional safety
is not considered in the model-based approach by Gomaa.
      </p>
      <p>
        Summary In order to provide information for deciding a product line
strategy, we see a potential to apply a model-based approach and in particular extend
the PLUS concept describe by Gomaa. The characteristics of the product line
can be described from di erent views which are necessary to capture functional
safety related attributes as well. The PLUS model proposed by Gomaa [
        <xref ref-type="bibr" rid="ref12">12</xref>
        ] has a
potential to be extended to cover both functional safety and product lines as well
as being extended to cover the systems engineering dimension. In the following
we present and discuss our approach.
3
3.1
      </p>
    </sec>
    <sec id="sec-3">
      <title>Approach</title>
      <sec id="sec-3-1">
        <title>General Idea</title>
        <p>
          In order to be able to take functional safety into consideration while planning
the product line, the relevant information need to be available already in early
development phases. Functional safety requires a holistic approach being able
to capture information throughout all development phases. On the one hand we
can build upon model-based product line engineering approaches as for example
PLUS [
          <xref ref-type="bibr" rid="ref12">12</xref>
          ]. On the other hand, model-based development is already common
for software development and therefore it is possible to build upon already
established practice. We aim to answer the research question: How can we add
functional safety related artifacts to a product line model?
        </p>
        <p>In Figure 1 we present the general concept of our approach. The model itself
is an Add-on to the PLUS approach of Gomaa taking functional safety into
consideration. The model-based approach we aim to develop shall contain both
development artifacts and safety-related artifacts. By the help of not just adding
separate diagrams for modeling the safety-critical functions, it is possible to
identify and capture dependencies between safety-related and non-safety-related
functions. When all information is captured in one model, automatic consistency
checks can be made to identify potential violations of safety goals in speci c
con gurations. Change requests shall be analyzed automatically and may result
in an impact analysis report extracted from the model. Since the main goal of our
work is to derive a safety case for each product variant, the model shall enable
the automatic generation of all necessary documents, i.e. the safety case, for a
speci c product variant. This can be realized by using a product con guration as
an input. Prede ned internal rules may extract the relevant information from the
model and create the required documentation. A model that captures all relevant
information will enable future product line instantiations and evolution.</p>
        <p>We developed a model-based approach for the concept phase capturing
commonality and variability on the one hand and the ISO 26262 related
information on the other hand. We applied our approach using a steer-by-wire example
(Comfort Drive Control - CDC) from the construction equipment domain.</p>
        <p>Each machine has a mechanical steering wheel, but a steer-by-wire solution
can be ordered as an option. We foresee two possible variants for the CDC - a)
left-right steering, which imitates the steering wheel functionality using a lever
and b) joystick steering, which adds forward/backward movement to the
leftright steering. The joystick steering has a higher criticality in comparison to the
left/right steering, since the required communication to the engine and gearbox
may fail with less possibility to control.
3.2</p>
      </sec>
      <sec id="sec-3-2">
        <title>Approach - Concept Phase</title>
        <p>The main challenge we identi ed is the actual mapping of the V-model-based
process described in the ISO26262 to the product line development process. For
the concept phase the standard requires that the safety critical features (items)
are identi ed, a hazard analysis is performed in order to identify the criticality
of the features and applicable safety concepts shall be de ned. The standard
furthermore requires that di erent concepts are analyzed and evaluated to choose
the appropriate safety concepts.</p>
        <p>
          Process
During the concept phase the product line strategy is derived that speci es which
reusable functions are to be implemented in a platform and which functions are
product speci c and will be developed in the application engineering phase.
Generally, it needs to be decided how the items and concepts are mapped on the
common platform or the speci c applications. We furthermore aim to capture
the variability for items and safety concepts to enable the correct implementation
at later design stages. We utilize Use Case diagrams and Feature Diagrams from
PLUS for the concept phase and add additional safety related properties. The
activity Product Line Analysis initializes the concept phase and information
about the targeted products, the demanded features and which existing technical
solutions shall be reused are collected and provided for further analysis. The
use cases are collected in the Use Case Diagram and the required features are
derived and documented in the Feature Diagram. As a Hazard and Risk Analysis
(HARA) [
          <xref ref-type="bibr" rid="ref1">1</xref>
          ] we perform a Preliminary Hazard Analysis (PHA) [
          <xref ref-type="bibr" rid="ref13">13</xref>
          ] and the
information from the diagrams are used for the hazard analysis. A model-based
approach to document a PHA has been presented in [
          <xref ref-type="bibr" rid="ref14">14</xref>
          ] and has not yet been
explored in our work. Today the PHA is documented in a separate table. After
performing the PHA, the resulting hazards, Automotive Safety Integrity Level
(ASIL), risk reduction strategies and operational constraints are added to the
diagrams. This information will be used for later development stages. The results
of all analysis are fed back to the Product Line Analysis step to review, adapt
and improve the product line concept.
        </p>
        <p>Process - Use Case Diagram
The usage of the machines and relation between the machine functionality and
the operators or bystanders are captured in the Use Case Diagram. Apart from
the variability notation proposed in the PLUS model, we introduce a unique
title for each scenario and add the stereotypes «hazard», «mitigation strategy»
and «operating constraints». In Figure 2 our approach is applied to the
steerby-wire example. The di erent operating modes need to be de ned and in this
example we visualize the scenario pallet handling. In other scenarios as idling
or maintenance where the CDC is also involved di erent hazards are related.
The «hazard» documents the hazards identi ed in the PHA and the related
ASIL are added as a property. For the optional use case Left/Right Steering
the hazard Unintended steering is connected. There are two variants of this
hazard, while for product group 1 the hazard has an ASIL A, for product group
2 an ASIL B is identi ed. For Forward/Backward Movement we connect the
hazard Unintended Forward/Backward Driving with an ASIL D. From the last
generation of the product in which only Left/Right Steering was implemented,
a mitigation strategy can be reused for this use case. A mitigation strategy
is a possibility to reduce the criticality of the item. In this case the strategy
«Independent Monitoring of Outputs» was used in an earlier generation. The
property Reduction ASIL is re ecting on how the ASIL could be reduced with
the help of the mitigation strategy. The stereotype «operation constraint» can
be applied, when knowledge about constraints are available. In this case the
activation of the CDC shall only be allowed in o -road usage of the machine to
reduce the probability of accidents. By the help of such constraints the hazards
related to application during on-road usage can be excluded. The related hazards
and mitigation strategies are supporting practitioners to take the safety-related
information into consideration when designing the system.</p>
        <p>
          Process - Feature Diagram
The Feature Diagram of Gomaa captures di erent types of variability. We add
the stereotypes «hazard» and «safety concept» and di erent dependencies types.
We applied the feature model to the steer-by-wire case in Figure 3. The product
family "Machine Type X" is having the «common feature» "Vehicle Movement".
The feature "Vehicle Movement" is in each machine represented by the «common
feature» "Steering by Steering Wheel". It is optional that the "CDC Function" is
used. The feature group "CDC steering" can be represented by only one fo the
subfeatures "Lever Control" with ASIL A, "Lever Control" with ASIL B or the
"Joystick Control". The feature "Lever Control" with ASIL A is related to the
Left/Right Steering use case in the Use Case Diagram. There di erent hazards
have been identi ed for two di erent product groups. The hazards related to
these groups are related to the two di erent features for "Lever Control". The
leaves of the safety critical features are getting the attributes ASIL and
SafetyFeature. So for example the feature "Joystick Control" is a "Safety-Feature" and
the ASIL D has been identi ed in the PHA. The feature group "Independent
Monitoring" is grouping the mitigation strategy features "Lever Monitoring" and
"Joystick Monitoring". These features are identi ed by the attribute "Safety
Concept" and the possible reduction of the ASIL. Furthermore the mitigation
strategy may add new hazards which are represented by the ASIL level. We utilize
the dependencies proposed in [
          <xref ref-type="bibr" rid="ref15">15</xref>
          ] and more speci c «synergetic» to show that
features shall be implemented to work in parallel with regular synchronization.
In our case the "CDC steering" shall be monitored by the feature "Independend
Monitoring". In later stages of the development this dependency can be re ned
by adding the maximal monitoring intervals. By the help of the «excluded»
dependency the con guration constraints are captured. It is not allowed that there
is a machine that has a "Joystick Control" which is monitored by the feature
"Lever Monitoring".
4
        </p>
      </sec>
    </sec>
    <sec id="sec-4">
      <title>Discussion</title>
      <p>The presented approach focuses on the concept phase and to manage functional
safety in product lines. We use the PLUS notation and add safety-related
stereotypes to the Use Case Diagram and Feature Diagram. In the following we discuss
how our approach helps to overcome some of the challenges.</p>
      <p>1) Aid documenting safety concepts in a PL: Documenting safety
concepts and taking variability into consideration is important and we document
the safety concepts and their variability as well as exploring the dependencies
between features and safety concepts. This may aid practitioners in designing
the product line.</p>
      <p>2) Support in extracting a safety case for each con guration: Part
3 of the ISO 26262 and speci cally the requirement 5.4.1 guided us which
information is required to be documented for an item. We added the required
properties as new stereotypes and added relations. Rules and templates need to
be developed to proof the extraction of information, which is part of our future
work.</p>
      <p>3) Support in choosing a product line concept: When moving towards
a product line, a product line concept needs to be chosen. This concept de nes
which features should be provided by a common platform and which features
are product speci c. By providing information about variability and functional
safety in our approach, the development team can make informed decisions.</p>
      <p>4) Support in PL instantiation: By having a model when a new product
is planned, rework may be avoided because all details are stored in one model.
Furthermore having knowledge about related hazards supports the
understanding of product line. An impact analysis of projected changes can be supported
by model-internal analyses, which is not yet implemented. This will improve
understanding the impact of the change and help identifying the a ected parts of
the product line.</p>
      <p>5) Flawed Communication is a threat to the successful development of
safety critical products. A model-based approach helps to create a common view
on the one hand and support a better understanding on the other hand, when
the speci ed solutions can be simulated. Since we also combine functional safety,
dependencies and variability, relevant information is not hidden anymore as it is
the case in a document-based approach.
5</p>
    </sec>
    <sec id="sec-5">
      <title>Conclusion</title>
      <p>
        In this work we have investigated how functional safety can be managed during
concept phase in industrial product lines. We identi ed that model-based
development concepts have a potential to aid product line engineering and support
focusing on functional safety at the same time. We propose a model-based
approach for the concept phase de ned by the ISO 26262 which is based on the
PLUS approach proposed by Gomaa [
        <xref ref-type="bibr" rid="ref12">12</xref>
        ]. We use the Use Case Diagram and
Feature Diagram to capture product speci c properties and functional safety at
the same time. We applied our approach to an industry related steer-by-wire
example, visualizing the applicability of our approach. In section 4 we discussed
how our approach helps to overcome the challenges of managing functional safety
in product lines. In the scope of this paper we did not focus on performing
consistency checks in the model. By enhancing the PLUS model with information
about hazards, safety mechanisms, safety-related features and dependencies
between features such consistency checks become possible and necessary because
of the growing complexity of such models.
      </p>
      <p>
        The research presented by Lee et al. [
        <xref ref-type="bibr" rid="ref15">15</xref>
        ] shows the possibilities of performing
consistency checks for feature models that include dependencies. Further research
is necessary to perform consistency checks with a functional safety focus taking
the hazards and safety mechanisms into consideration as well. It is also necessary
to map to other development stages from the standard to the product line process
and explore the impact of product line strategies more in detail.
      </p>
      <p>It is furthermore possible to extend the model-based approach by a state chart
diagram where machine states, potential hazards and safe states documented.
This can be useful for later development stages when machine states are re ned.</p>
    </sec>
    <sec id="sec-6">
      <title>Acknowledgments</title>
      <p>The research leading to these results has received funding from the ARTEMIS
Joint Undertaking under grant agreements no269265 and no295373, Vinnova and
the KKS-funded ITS- EASY Post Graduate School for Embedded Software and
Systems.</p>
    </sec>
  </body>
  <back>
    <ref-list>
      <ref id="ref1">
        <mixed-citation>
          1. ISO:
          <article-title>ISO 26262 Road vehicles { Functional safety (</article-title>
          <year>2011</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref2">
        <mixed-citation>
          2.
          <string-name>
            <surname>Bosch</surname>
          </string-name>
          , J.:
          <article-title>Maturity and evolution in software product lines: Approaches, artefacts and organization</article-title>
          .
          <source>Software Product Lines</source>
          (
          <year>2002</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref3">
        <mixed-citation>
          3.
          <string-name>
            <surname>Lutz</surname>
            ,
            <given-names>R.R.</given-names>
          </string-name>
          :
          <article-title>Analyzing software requirements errors in safety-critical, embedded systems</article-title>
          . In: Requirements Engineering,
          <year>1993</year>
          ., Proceedings of IEEE International Symposium on,
          <source>IEEE</source>
          (
          <year>1993</year>
          )
          <volume>126</volume>
          {
          <fpage>133</fpage>
        </mixed-citation>
      </ref>
      <ref id="ref4">
        <mixed-citation>
          4.
          <string-name>
            <surname>Chen</surname>
            ,
            <given-names>D.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Johansson</surname>
            ,
            <given-names>R.</given-names>
          </string-name>
          , Lonn, H.,
          <string-name>
            <surname>Papadopoulos</surname>
            ,
            <given-names>Y.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Sandberg</surname>
            ,
            <given-names>A.</given-names>
          </string-name>
          , Torner,
          <string-name>
            <surname>F.</surname>
          </string-name>
          , Torngren, M.:
          <article-title>Modelling support for design of safety-critical automotive embedded systems</article-title>
          . In: Computer Safety, Reliability, and Security. Springer (
          <year>2008</year>
          )
          <volume>72</volume>
          {
          <fpage>85</fpage>
        </mixed-citation>
      </ref>
      <ref id="ref5">
        <mixed-citation>
          5.
          <string-name>
            <surname>Oertel</surname>
            ,
            <given-names>M.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Schulze</surname>
            ,
            <given-names>M.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Peikenkamp</surname>
            ,
            <given-names>T.</given-names>
          </string-name>
          :
          <article-title>Reusing a functional safety concept in variable system architectures</article-title>
          .
          <source>In: 7th International Workshop on Model-Based Architecting and Construction of Embedded Systems</source>
          . (
          <year>2014</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref6">
        <mixed-citation>
          6.
          <string-name>
            <surname>Sun</surname>
            ,
            <given-names>H.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Hauptman</surname>
            ,
            <given-names>M.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Lutz</surname>
            ,
            <given-names>R.</given-names>
          </string-name>
          :
          <article-title>Integrating product-line fault tree analysis into aadl models</article-title>
          .
          <source>In: High Assurance Systems Engineering Symposium</source>
          ,
          <year>2007</year>
          . HASE '
          <volume>07</volume>
          . 10th IEEE. (
          <year>Nov 2007</year>
          )
          <volume>15</volume>
          {
          <fpage>22</fpage>
        </mixed-citation>
      </ref>
      <ref id="ref7">
        <mixed-citation>
          7.
          <string-name>
            <surname>Montecchi</surname>
            ,
            <given-names>L.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Lollini</surname>
            ,
            <given-names>P.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Bondavalli</surname>
            ,
            <given-names>A.</given-names>
          </string-name>
          :
          <article-title>Dependability concerns in modeldriven engineering</article-title>
          . In: Object/Component/Service-Oriented
          <string-name>
            <surname>Real-Time Distributed Computing Workshops (ISORCW)</surname>
          </string-name>
          ,
          <year>2011</year>
          14th IEEE International Symposium on.
          <source>(March</source>
          <year>2011</year>
          )
          <volume>254</volume>
          {
          <fpage>263</fpage>
        </mixed-citation>
      </ref>
      <ref id="ref8">
        <mixed-citation>
          8.
          <string-name>
            <surname>Kaiser</surname>
            ,
            <given-names>B.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Liggesmeyer</surname>
            ,
            <given-names>P.</given-names>
          </string-name>
          , Mackel, O.:
          <article-title>A new component concept for fault trees</article-title>
          .
          <source>In: Proceedings of the 8th Australian Workshop on Safety Critical Systems and Software - Volume 33. SCS '03</source>
          ,
          <string-name>
            <surname>Darlinghurst</surname>
          </string-name>
          , Australia, Australia, Australian Computer Society, Inc. (
          <year>2003</year>
          )
          <volume>37</volume>
          {
          <fpage>46</fpage>
        </mixed-citation>
      </ref>
      <ref id="ref9">
        <mixed-citation>
          9.
          <string-name>
            <surname>Gomez</surname>
            ,
            <given-names>C.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Liggesmeyer</surname>
            ,
            <given-names>P.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Sutor</surname>
            ,
            <given-names>A.</given-names>
          </string-name>
          :
          <article-title>Variability management of safety and reliability models: An intermediate model towards systematic reuse of component fault trees</article-title>
          . In Schoitsch, E., ed.: Computer Safety, Reliability, and
          <string-name>
            <surname>Security</surname>
          </string-name>
          . Volume
          <volume>6351</volume>
          of Lecture Notes in Computer Science., Springer Berlin Heidelberg (
          <year>2010</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref10">
        <mixed-citation>
          10.
          <string-name>
            <surname>Biggs</surname>
            ,
            <given-names>G.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Sakamoto</surname>
            ,
            <given-names>T.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Kotoku</surname>
            ,
            <given-names>T.</given-names>
          </string-name>
          :
          <article-title>A pro le and tool for modelling safety information with design information in sysml</article-title>
          .
          <source>Software &amp; Systems Modeling</source>
          (
          <year>2014</year>
          )
          <volume>1</volume>
          {
          <fpage>32</fpage>
        </mixed-citation>
      </ref>
      <ref id="ref11">
        <mixed-citation>
          11.
          <string-name>
            <surname>Liu</surname>
            ,
            <given-names>J.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Dehlinger</surname>
            ,
            <given-names>J.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Lutz</surname>
          </string-name>
          , R.:
          <article-title>Safety analysis of software product lines using state-based modeling</article-title>
          .
          <source>Journal of Systems and Software</source>
          (
          <year>2007</year>
          )
          <year>1879</year>
          {
          <fpage>1892</fpage>
        </mixed-citation>
      </ref>
      <ref id="ref12">
        <mixed-citation>
          12.
          <string-name>
            <surname>Gomaa</surname>
          </string-name>
          , H.:
          <article-title>Designing software product lines with UML</article-title>
          .
          <string-name>
            <surname>Addison-Wesley</surname>
            <given-names>Boston</given-names>
          </string-name>
          , USA (
          <year>2004</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref13">
        <mixed-citation>
          13.
          <string-name>
            <surname>Ericson</surname>
            ,
            <given-names>C.</given-names>
          </string-name>
          :
          <article-title>Hazard analysis techniques for system safety</article-title>
          . Wiley-Interscience (
          <year>2005</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref14">
        <mixed-citation>
          14.
          <string-name>
            <surname>Marielle</surname>
            ,
            <given-names>P.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Thomas</surname>
            ,
            <given-names>F.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Belmonte</surname>
            ,
            <given-names>F.</given-names>
          </string-name>
          :
          <article-title>Interoperability between risk assessment and system design for railway safety critical signalling system development</article-title>
          . In: 17eme Congres de Maitrise des Risques et de Surete de Fonctionnement,
          <source>IMDR</source>
          (
          <year>2010</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref15">
        <mixed-citation>
          15.
          <string-name>
            <surname>Lee</surname>
            ,
            <given-names>Y.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Yang</surname>
            ,
            <given-names>C.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Zhu</surname>
            ,
            <given-names>C.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Zhao</surname>
            ,
            <given-names>W.:</given-names>
          </string-name>
          <article-title>An approach to managing feature dependencies for product releasing in software product lines</article-title>
          .
          <source>In: Reuse of O -the-Shelf Components</source>
          . Springer (
          <year>2006</year>
          )
          <volume>127</volume>
          {
          <fpage>141</fpage>
        </mixed-citation>
      </ref>
    </ref-list>
  </back>
</article>