Impact and Overrun Detection in Automated Rail Transport: Technologies and Data Foundations
Impact and overrun detection are important complements to vehicle-based environmental perception systems for automated train operation.
In order to develop fully automated, driverless operation, sensor-based environmental perception systems are required to perform track monitoring. To date, most projects have primarily relied on systems (camera, radar and Lidar) that detect obstacles but cannot reliably identify impact or overrun events. For fully automated operation on open mainline networks, a system is therefore required that not only detects obstacles and initiates braking decisions, but also reliably detects impact and overrun events and triggers appropriate follow-up measures.
Within the sector initiative Digitale Schiene Deutschland, DB InfraGO AG has conducted several projects in the fields of environmental perception as well as impact and overrun detection. The work commenced with a feasibility study commissioned in 2019 to the IFB – Institute of Railway Engineering, the German Aerospace Center (DLR) and the University of Stuttgart. Subsequently, two prototype development projects were initiated: one focusing on overrun detection and one on impact detection. Based on the insights gained, a combined and comprehensive project entitled “AI Methods for Condition Monitoring and Demand-Oriented Maintenance of Railway Vehicle Structures (KI-MeZIS)/KI-Methoden in der Zustandsüberwachung und bedarfsangepassten Instandhaltung von Schienenfahrzeugstrukturen (KI-MeZIS)” was ultimately launched.
The system was developed within these projects up to Technology Readiness Level 6. The work has now been completed. It has become apparent that, for many research questions, relevant data were collected and analysed for the first time. The results of the impact and overrun detection projects are presented in a recent technical article published in the November issue of Eisenbahntechnische Rundschau (only in German).
