Accurate prediction of train arrival times helps increase efficiency across the railway sector and for connected stakeholders, while enabling better utilisation of resources.

Building on previous phases, the RNE ETA Programme has enhanced ETA calculation by allowing stakeholders to incorporate multiple data sources and by introducing an AI-based algorithm.

Tests confirmed that ETAs calculated with artificial intelligence are more precise than those based on linear time-shifting and the previously used estimates in RNE’s Train Information System (TIS). Since November 2025, the new AI-based Network ETA has been implemented and made available in TIS, covering the entire European network through the system’s Network ETA module. This development also makes TIS more attractive for additional stakeholders, such as terminals and CTOs, to connect and exchange train-related data.

In parallel, work on exchanging ETA accuracy via an adapted Telematic TSI message has been progressing. Newly developed accuracy identifiers allow the sector to select the most reliable forecasts and monitor improvements over time.

Further enhancements will be addressed through the ETA Preference and ETA Contingency concepts – definition and implementation planned from 2026 onwards.