An IMU as the basic vehicle-borne sensor within a modified Multi-Hypothesis Based Map-Matching Algorithm for Train Positioning

Abstract

For certain types of railroad lines replacing the equipment for precise train positioning along the track by suitable low-cost sensors and a digital map on the train can result in a more cost-efficient railway operation. This paper presents a modular multi-hypothesis based map-matching (MHMM) approach providing track-selective localization of the train by fusing the data of several train-borne positioning sensors. The MHMM is based on four main process steps: hypothesis generation, update, evaluation, and selection. These are periodically run through in a cycle. While previously presented contributions examined various sensor data for evaluating the different position hypotheses, especially right after passing a switch facing, this work focuses on the step of the position hypotheses update. Here, the inertial measurement unit (IMU) as the central navigation sensor is analyzed. First, its acceleration data are fused with speed observations from additional sensors, e.g. a GNSS receiver, providing precise information about the travelled distance. Then, the thus updated relative position on the track is adjusted by evaluating the track’s curvature, fusing the IMU yaw rate measurement with heading observations. Both data fusion steps are based on Kalman filters in order to calibrate the IMU. Compared to solely determining the travelled distance by an odometer or a Doppler-radar, the IMU-based hypotheses update provides a more accurate position information that builds the basis for the following hypotheses evaluation step within the MHMM. The second main benefit of the thus modified algorithm is its higher cycle frequency.