SWITONIC

Background

SWITONIC was an industry collaboration project funded by NCCR Automation, running from October 2023 to September 2025. On the academic side, the project was led by the University of Applied Sciences and Arts Northwestern Switzerland (FHNW) in collaboration with ETH Zurich. Industry partners included the signalling team of Stadler and the train operator Schweizerische Südostbahn (SOB).

The objective of SWITONIC was to deploy advanced optimization algorithms within active driver assistance systems of real trains and benchmark predictive control against common industry approaches under realistic operating conditions. The results demonstrated the practical viability and performance advantages of model-based predictive control in rail applications.

SWITONIC formed the technological foundation for nuorail. Following the encouraging mid-project results, the concept for nuorail was developed at FHNW and the company was subsequently incorporated to further translate the research into scalable industrial solutions.

Project at a glance

• Algorithm tested on a Stadler FLIRT 3 EMU (4-car unit)
• Test track from Biberbrugg to Steinerberg, with intermediate stops at Altmatt, Rothenthurm and Sattel (34 km round-trip)
• 3 night test and 2 day test campaigns, the former for accurate energy consumption estimation (green wave) and the latter for validation in combination with supervising modules
• Predictive controller running onboard interfaced with Stadler's driver assistance software (see here for more details on our architecture)
• Reproducible results confirmed energy consumption reductions and improved punctuality

An example from the tests

The figure on the right shows the speed profile (velocity over distance) from one of the tests, together with speed limits and the altitude profile (gray shaded area). The trajectory clearly shows how the algorithm leverages the future slope in order to apply as little traction as necessary and arrive on time.