Research on a hybrid map matching algorithm for Global Navigation Satellite System based train positioning

GNSS has been proved to have great potential for Safety-of-Life critical rail applications, particularly the train control technique and railway signalling. In the GNSS based train positioning scheme, although with the aid of inertial sensors (e.g. the odometer, gyro, accelerator and Doppler radar) some systematic and random errors could be reduced or limited by an appropriate measuring method and data fusion filtering, it is significant to improve and guarantee the positioning precision and integrity performance by using the map matching (MM) technique in a cost effective way. In this paper, the structure of an electrical track map database is designed according to the requirements of precision and efficiency, the architecture of a GNSS based train positioning system integrating INS sensors is introduced, and a novel hybrid map matching algorithm is proposed, in which the determined train position is the integration of the position solution from multi-sensor fusion, the identification of the similarity or matching probability, and heading validation, with different track map levels. As the “point-to-curve” and “point-to-point” matching strategy are adopted with the provided feature of track map data, the adaptive performance and completeness of the map matching algorithm is guaranteed and improved. A field test in the Qinghai-Tibet line demonstrates that the proposed algorithm earns high position decision accuracy and integrity with simple implementation, which is of great practical value to precise train control and railway signalling.