A Knowledge Based Approach for Accident Analysis: Application to the Safety of Automated Rail Transport Systems

A Knowledge Based Approach for Accident Analysis: Application to the Safety of Automated Rail Transport Systems

Lassad Mejri
Copyright: © 2018 |Pages: 16
DOI: 10.4018/IJIRR.2018100104
(Individual Articles)
No Current Special Offers


This article introduces a research aiming at the development of a decision support system concerning the approval of automated railway transportation systems. The objective is to evaluate the degree of compliance of the transportation system according to the security standards by the simulation of the scenarios of accident. To reach this target, the authors envisaged an approach Rex (Return of experience) who draws lessons of accidents lived and/or imagined by the experts of the analysis of security in the NRITS (National Research Institute on Transports and their Security): currently IFSTAAR. The approach consists in offering aid to the experts by a reuse of the accidents already validated. This approach provides to the experts a class of similar accidents situations to the new case. The case-based reasoning is then exploited allowing choosing one under group of historical cases that can help in the resolution of the new case introduced by the experts.
Article Preview

State Of The Art, Goals And Research Orientations

The problem of the analysis of safety could be registered within the framework of the resolution of problems generally and in the field of Rex (Return of EXperience) in particular. According to (Richard 1997), a problem can be defined by: an initial state, an objective state and obstacles. In the same way, an accident in the automated rail transportation is characterized by an initial situation of the system, intermediate states before the accident and finally a critical situation which characterizes the accident. We browze now briefly the typology of the problems such as defined in literature of Problem Solving (Poissant, Poellhuber, & Falardeau, 1994; Pierret-Golbreich, 1998).

The most famous classification distinguishes Structured and not Structured problems: In the structured problems, all the data are known in advance, rules are precise and clear, a continuation of operations leads with certainty to a single possible solution. These problems could be modeled. In the not structured problems, the data are not quite known in a definitive way, new data is introduced into the course of the problem processing; rules can change from a situation to another. It can have various paths which allow reaching the solution. Such problems could be modeled with difficulty.

The problem of the security analysis in the automated transport is a not structured problem because the data bound to an accident are not well known in advance. Several human and automatic actors are involved in an accident.

The acquisition of knowledge strikes obstacles to extract the approach of the experts of the security analysis. Indeed, their approach remains implicit, intuitive and not easily able to be formalized. We leave then the classic approach of knowledge extraction towards a framework of knowledge engineering. The research works in the knowledge engineering aim at modeling three types of knowledge (Uschold & Tate, 1998), (Charlet, 2002): the knowledge of the domain, the tasks and the methods. All these Knowledge form, in knowledge engineering, the conceptual model. The analysis of the Top-Down and Bottom-Up engineering approaches (Van Heijst, Schreiber, & Wielenga, 1997), (Schreiber et al., 2000), (Teulier, Charlet, & Chounikine, 2006) such as KOD/KADS/COMMON-KADS/MACAO/MACAOII (Aussenac-Gilles, 2006), (Lépine & Aussenac-Gilles, 1996), etc., directed us to a pragmatic approach:

Complete Article List

Search this Journal:
Volume 14: 1 Issue (2024): Forthcoming, Available for Pre-Order
Volume 13: 1 Issue (2023)
Volume 12: 4 Issues (2022): 3 Released, 1 Forthcoming
Volume 11: 4 Issues (2021)
Volume 10: 4 Issues (2020)
Volume 9: 4 Issues (2019)
Volume 8: 4 Issues (2018)
Volume 7: 4 Issues (2017)
Volume 6: 4 Issues (2016)
Volume 5: 4 Issues (2015)
Volume 4: 4 Issues (2014)
Volume 3: 4 Issues (2013)
Volume 2: 4 Issues (2012)
Volume 1: 4 Issues (2011)
View Complete Journal Contents Listing