An Ontological Approach to Support Dysfunctional Analysis for Railway Systems Design

Abstract

Dysfunctional analysis is an essential and demanding task in the early development stages of safety-critical systems (SCSs). Nevertheless, current practices present several drawbacks. Generally, a common dysfunctional analysis conceptualization is missing and it is dependent on safety analysis techniques. Moreover, some safety analysis methods require well-known system behaviors expressed by dynamic models such as sequence diagrams and finite automata. However, the dynamic character of these models increases their susceptibility to changes and then they are not obtainable in the early design stages. Since dysfunctional analysis highly relies on the experience of safety analysts and the feedback (REX) obtained from previous systems development, there is a need to formalize this knowledge domain in a structured way to ensure its future reuse. Furthermore, safety measures derived from this dysfunctional analysis approach must be strongly linked to a goal-oriented perspective and adapted to a specific context. For this purpose, this paper presents a real-world semantics interpretation and conceptualization of dysfunctional analysis related concepts based on the Unified Foundational Ontology (UFO) and well-known standards to avoid ambiguities. The proposed Dysfunctional Analysis Ontology (DAO) aims to provide a systematization of the goal-oriented dysfunctional analysis through a terminological clarification in order to prevent hazards in the first design phases. Then, a DAO formalization is proposed using the Web Ontology Language (OWL). Finally, the DAO pattern is applied to two different real critical scenarios from the railway domain in order to illustrate and evaluate this ontological approach.

Introduction to the Special Issue on Software-Intensive Autonomous Systems: Methods and applications

“…With the aim to provide a conceptualization of dysfunctional analysis, a reference domain ontology called DAO (Dysfunctional Analysis Ontology) was previously developed [4]. DAO is grounded on Unified Foundational Ontology (UFO) which is an upper-level ontology [5].…”

“…The formalisation of this semantic link between safety measures and safety goals is crucial since it improves the safety assurance and hazards mitigation. Further details about DAO and GOSMO development process may be found, respectively in [4] and [7]. In the present study, DAO and GOSMO are used and combined with other models to have a structured safety model-based process.…”

“…In this work, the representation is based on UML (Unified Modeling Language) [14], a semi-formal, enrichable and structured language. The modelling task is structured around the expertise knowledge and competency questions, and based on semantic formalisms, transformation rules and frameworks for transition from one model to another [4]. Indeed, the MBE can ensure the traceability of business and safety requirements as described in our proposal.…”

“…The latter is a UML profile for conceptual modeling and it incorporates foundational distinctions defined in UFO. The interpretation of failure and its related concepts in real-world semantics may be found in [4].…”

An Ontology of Security from a Risk Treatment Perspective

“…So CSO does not seem to commit to UFO in its full extent. The Dysfunctional Analysis Ontology (DAO) [8] continues the Goal-Oriented Safety Management Ontology (GOSMO) and aims at providing a systematization of the goal-oriented dysfunctional analysis through a terminological clarification in order to prevent hazards. They are represented in UML and OWL, making the same (in our view, mistaken) choice of interpreting Safety Measures as an Actions.…”