|Future railway applications, such as train-side collision avoidance, virtual coupling or autonomous train driving demand reliable and accurate train localization. We focus on exclusive onboard train localization without additional way-side infrastructure. Common approaches for onboard train localization are based on measurements of a global satellite navigation system (GNSS). Well-known methods for improved reliability and accuracy are the combination with inertial navigation systems (INS) or dead reckoning. However, these approaches show an unbounded growth of an error for longer GNSS outages. Reliable train localization requires a long-term stability with bounded errors in long tunnels, at underground stations or in long stretches of dense forests. This paper presents two methods as well as a new system concept for onboard train localization with GNSS redundancy and long-term stability. The first method is signature based train localization approach from independent IMU and magnetometer measurements. The second method introduces an independent speed estimation approach without a wheel speed sensor nor a GNSS. A track signature is considered here as a signal of a track feature which is sampled over the 1-D positions of a track. It is possible to estimate a location from a comparison of the latest measured signatures and reference signatures from a map. This paper contains further an analysis of the signatures on straight tracks, parallel tracks, a switch, and a curve. The used signatures comprise eight signatures from magnetic, curvatures, attitude and vibration. All signatures use recorded data from real train runs of a regional train.
Proceedings of the 30th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2017)
September 25 - 29, 2017
Oregon Convention Center
|3231 - 3237
|Cite this article:
Heirich, Oliver, Siebler, Benjamin, "Onboard Train Localization with Track Signatures: Towards GNSS Redundancy," Proceedings of the 30th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2017), Portland, Oregon, September 2017, pp. 3231-3237.
ION Members/Non-Members: 1 Download Credit