Multi-sensor fusion for seamless navigation in railway domain.

Abstract

The European Union (EU) aims at making railway a more attractive transportation method by improving its efficiency and reducing its costs. These achievements could be covered with the migration from ETCS level 2 to ETCS level 3. Many projects related to new positioning systems have been funded by The European Union. Most of these positioning systems are based on GNSS, due to the key role that GNSS will play in the migration to ETCS level 3. One of the problems on using only GNSS systems is the lack of availability of them. During railway operation, there are areas with potential GNSS outages, such as urban canyons, woods or other possible signal blockers and disturbances. Moreover, it is a fact that GNSS signals are not reachable, nor reliable in tunnels or indoor environments. For GNSS to be able to have a key role in the next years in railway security, the afore mentioned lack of availability has to be solved. To cope with this issue, a multi-sensor approach with software enhancements is proposed in this dissertation. The objective of this research work deals with fusing different sensors and creating new software strategies to achieve a higher availability with the best possible accuracy. The seamless position will benefit in all the operation modes, from the train station to a harsh environment for satellites, during the train operation. The scope of the dissertation is to create a multi-sensor positioning system including GNSS, Inertial Measurement Unit (IMU), and Ultra Wide Band (UWB) with other software techniques to obtain a position estimation with a 100% availability for railway systems. This work shows the different steps from the study of the state of the art, going through the implementation, and ending with the performance analysis of the algorithm developed. This research work has been conducted under different European projects such as ERSAT-EAV, FR8RAIL or X2Rail-2, in which CEIT has participated.