Efficient formalization of railway interlocking data in RailML

Railways wish to reduce the costs of engineering interlocking systems by simplifying the exchange of technical information between stakeholders. Exchanging interlocking data in a machine-readable and standardized format removes inefficiency due to misinterpretation and conversion of data from non-standard, often paper-based formats. This paper proposes an UML class representation of the interlocking data that maps the characteristics of the interlocking system. These data model represents both tabular and geographical interlocking. The model reflects the topology of the railway network that the interlocking controls, routes, relations between signal aspects, speed indication signals, automatic train protection (ATP) and the state of movable and non-movable track elements. The data are machine-readable RailML, an incarnation of XML that is gaining the status of standard in railway modelling. We applied our approach to model the Dutch Santpoort Noord station, starting from paper-based track and signal plans. The resulting database captures all the features of the interlocking, whilst removing data redundancy. The model has general validity and it is easily extendible to other interlocking and ATP systems. Our approach contributes to create standard, electronic interlocking databases, directly from technical documents currently used in the practical world. •We propose a database of railway interlocking mapping logical and physical features.•The database builds on a novel UML formalization of interlocking elements.•The standard machine-readable language RailML is used to structure the database.•Our formalization overcomes the limits of previous approaches proposed so far.•Using a standard database eases the communication process among stakeholders.