Generation and Evaluation of the Track Map Database for GNSS-based Train Positioning Using a Map-tool-chain

The train positioning technology based on GNSS (Global Navigation Satellite System) has greatly developed in the train control system because of its stability, low-cost, easy maintenance. The train positioning technology is used to prevent possible potential train-to-train collision, and it can also be utilized for automatic train operation and train scheduling. However, due to some inevitable defects in GNSS, such as SIS (Signal in Space) unavailability, multi-path effects and electromagnetic interference and so on, map-assisted train positioning technique is proposed to achieve a better performance. Map matching plays a crucial role in map-assisted train positioning, which uses position and heading data from positioning systems including GNSS, dead reckoning (DR) system and inertial navigation system (INS) to integrate and then compare with the track map database. The map matching algorithm finally distinguishes the actual track the train is running on between parallel tracks and matches the measured position to the right point on map route. So a digital track map database is indispensable for map-assisted GNSS-based applications. The track map database can be built based on the original measurements with high-accuracy GPS receivers. And developing a dedicated tool to assist map data collection and processing in field operation is of great importance, which can reduce the cost and workload considerably. In this paper, the requirements and necessities of digital track map database in train positioning are explained. The map-tool-chain is implemented, which consists of Station Schematic Editor, Station Map Data Recorder, Map Database Editor, and Map Database Validator. The four components serves the whole process of the generation of digital track map database, which includes four step: railway track surveying, data processing, map database generation and map database validation. There are a lot of methods using during the track map generation, such as outliers elimination in original measurement, uniform sampling, smoothness inspection in track curve and multiple GPS trajectories fusion. These algorithms will be elaborated in this paper. Furthermore, the map-tool-chain is used to generate a track map database for field applications in Hankou-Yichang Railway (HYR) and Jincheng Anthracite Mining Group in China. The results from field applications prove that the map-tool-chain can not only inspect and remove the outliers in raw data, but also make the generated track map better spacing and smoother. The proposed methods and map-tool-chain will be more competitive in developing rail applications based on GNSS positioning, including train control and supervision system, train collision early warning system and train integrity monitoring system.