ATO application scenarios – What tests should an automated train be able to pass?
Image above: Driverless journey in DLR's RailSET® driver's cab simulator.
In automated train operation (ATO), cameras and other sensors will take over the observation of the train's surroundings from the driver in higher degrees of automation (Grade of Automation - GoA 3/4). Artificial intelligence (AI) will also be used to process the sensor data. This poses a challenge for the necessary approval, as the safety of the new technical systems can no longer be proven by fully analysing how they work.
In the automotive sector, where such systems are also used for highly automated driving, the "scenario-based testing" approach has therefore been developed in recent years (see the PEGASUS project family). This involves systematically identifying the scenarios that occur in traffic and gradually deriving their possible variants and relevant parameterisations. The many concrete scenarios that emerge serve as tests for the technical system, most of which are executed simulation-based.
This approach is to be transferred to the railway sector in the ATO application scenarios project. Railway-specific operational design domains (ODD) are defined for which the respective authorisation should apply and which help to narrow down the required ATO scenarios. The derivation methodology and suitable means of describing railway scenarios are defined and demonstrated using an example ODD. Finally, software will be developed with which test cases can be generated as concrete parameterisations of scenarios, and its application will be demonstrated. The development of the reasoning for a certain completeness of the scenarios with regard to the ODD is also part of the project.
The activities are rounded off by researching project-relevant literature and a final validation of the project results to ensure that they are up to date, correct and suitable for the problem of approving ATO systems. While the project can certainly not deliver any final parameterised executable test cases for an ODD due to its limited scope, the aim is to develop a viable methodology for the scenario-based testing of railway ATO systems that can be used in practice.
Project title:
Systematisation of the application scenarios for ATO
Duration:
05/2024 to 04/2027
Related projects:
This project is managed by the department:
