Long Interval Of Large-scale Data-based Map-matching Techniques

The development of Intelligent Transportation System (ITS) has been one of the most important area in the field of transportation. Floating vehicle technology has been widely applied for traffic data collection in ITS. It adopts many advanced technologies, such as locating, communication, data storage, Geographical Information System (GIS) and so on. Vehicle location is one of the most important type of information in ITS, and Map-Matching is the primary way for obtaining accurate location information.First of all, this paper introduces the frame and key technologies of floating car system based on GPS/GIS, basic concept and theory of Map-Matching. Under the country tenth-five years development ITS plan"Hangzhou ITS traffic system demonstration project", this paper analyses the special Hangzhou traffic situation and puts forward a new compositive algorithm comparing the existing Map-Matching algorithms, and at last analyses the new algorithm from precision and real-time aspects.Concerning the large data, requirement of real-time and long interval of GPS data in floating car system, the author investigates the dedicated Map-Matching algorithms. This paper introduces the method of combining weight-based and map topology point to curve Map-Matching algorithms that fits the requirements of floating car system. The new Map-Matching algorithms can gain not only the ability of real-time matching in simple road situations but also the accuracy of results complicated situations respectively. Because of the long interval of GPS data, it's difficult to confirm there are no default paths in the matched paths, the Map-Matching algorithms contain a shortest path algorithm to process this situation and get an integrated vehicle path. The software is designed totally based on Object Oriented Architecture, which assures this system's robustness and, meanwhile, less expense of maintenance.The author also makes the analysis and evaluation about the performance of key algorithms used. The evaluation work is conducted with urban network in Beijing. This paper also demonstrates the feasibility and logical rationality of the design of Map-Matching algorithms by comparing it with other point to curve methods.Last but not the least, some suggestions are posed for future research.