Study On Digital Map Generation Method For Train Control System

With the development of CBTC (Communication Based Train Control) applications, train positioning as a key technology has been promoted greatly. Using the GPS and GIS information to get the accurate train position is a relative new method in this area.A train can use the database of stations and blocks together with GPS position readings to localize itself, but this all based on an accurate detailed digital map. High precision GPS track data could be used to create digital map, but it turns out too expensive and time-consuming. Therefore, how to use a large amount of low cost GPS data to create high precision train control digital map effectively and economically becomes significant and important in these applications.An algorithm named Multiple trajectory GPS data extraction and refinement was proposed in this paper to realize the automatic data refinement and updating of the digital map. High precision POI (Point of Interest) GPS data and a great number of low cost and less accurate track GPS data needed to be collected in this method. The POI position used as a restriction condition in calculating the track functions. This algorithm is composed by five parts, namely data collection, data pre-process, calculating track functions and results estimation respectively.A modified Hough transform was used as a data selector in the data pre-process part, the linear and curve functions was discussed respectively in the part of calculating the track functions. The thesis also present a number of evaluation metrics for estimating the quality of the maps the algorithm generated. The precision of the functions, the storage of the digital map and the speed of map matching proved that the algorithm is a feasible reasonable and economical method of creating train control digital map.At the end of the thesis, it is pointed that the algorithm could not only use in the train control system, but could be used in the ITS (Intelligent Traffic system) as a map generation method as well. Furthermore, it may be possible in the future for roadways not yet available within a GIS database to be appended using the information generated from the proposed method.