Train run forecasting helps to estimate a train’s arrival at a given location, which can significantly increase efficiency within the railway sector and connected stakeholders, and lead to a better utilisation of resources. The RNE ETA Programme, following the successful implementation of previous phases, aims to improve the accuracy of ETA by helping stakeholders to take different sources into account and introducing ETA calculation based on an algorithm using artificial intelligence. Tests have proven that ETAs calculated with the help of artificial intelligence were more accurate than ETAs based on linear time-shifting and current ETAs in the TIS. Implementing this new AI-based ETA into TIS makes it potentially more interesting for new stakeholders, such as terminals and CTOs, to connect to TIS and exchange train-related data.
Current activities are focused on the implementation of the so-called Network ETA covering the whole European network. Furthermore, the improvement of forecast accuracy is addressed through the ETA Preference concept, where the intention is to exchange on unplanned situations during the train run.
The newly developed accuracy identifiers enable the sector to select the most accurate forecasts and monitor forecast improvements related to a certain timeframe.