Bayesian train localization method extended by 3D geometric railway track observations from inertial sensors

The localization of trains in a railway network is necessary for train control or applications like autonomous train driving or collision avoidance systems. Train localization is safety critical and therefore the approach requires a robust, accurate and track selective localization. The proposed approach focuses on an onboard localization system using appropriate onboard sensors without any additional railway infrastructure. Global navigation satellite systems (GNSS) are very beneficial for this task, but the accuracy, measurement errors and the lack of availability in parts of the railway environment do not fulfill the requirements for a safety critical railway system. An inertial measurement unit (IMU) is used to increase robustness and accuracy, by sensing the position related effects of 3D track geometry on the train. In order to cope with multiple sensors, position related uncertainties and sensor errors, we propose a probabilistic approach with a Bayesian filter. In this paper we present a train localization approach using a particle filter in combination with multiple onboard train sensor measurements and a known track map. The particle filter estimates a topological position directly in the track map. First simulations show encouraging results in robustness and accuracy in critical railway scenarios.