Application And Research Of Map Matching Technique In Embedded GPS Navigation

In recent years, along with the extensive applications of the Geographic Information System, the Embedded System and the Satellite Positioning technique, the Intelligent Transportation System used to resolve the city transportation problem gets to develop quickly. As one of the important parts of the Intelligent Transportation System, the Vehicle Navigation System also has a rapid development, the Embedded Vehicle Navigation System based on GPS location technology as a new generation Vehicle Navigation System has already become a popular research direction in social researches. The map matching as a key technology of the Vehicle Navigation System, it has great impact and significance. its accuracy, real-time and adaptability will directly affect the Vehicle Navigation System's performance. This paper aim at the constraints of the embedded system in storage and computation, make a intensive study on the map matching technology. Through the three aspect analysis of the GPS abnormal signal, the data organization and management of digital map and the map matching algorithm, a complete map matching technology solutions is given at last, to ensure under the different road conditions, or in exceptional cases, the system both can has an accuracy and real time map matching result.The paper studies the theory of map matching technique and introduce the commonly-used map matching algorithm at present briefly, analyzed and compared their advantage and disadvantage. On this basis, specialize three important factors which have great impact to the map matching result:GPS positioning error, the organization of digital map data and map matching algorithm. On the study of the data organization and management of digital map, this paper give a hierarchical grid model of spatial data organization and a model storing road topology data in provincial to improve the speed of accessing the relevant road data in process of the map matching. Analyze the GPS positioning error and establish the GPS signal processing model, the GPS signal is divided into five kinds of states:initial state, normal state, low-speed state, abnormal state and invalid state, the GPS signal will be corrected in the different state. A comprehensive weight map matching algorithm which base on shortest distance, direction deviation and road topology is given, then establish the matching road selected model and matching points selected model. The complete map matching technology solutions is set up by the three process models and the detail introduction of the solution. The experiments are carried out to test correctness, timeliness and feasibility of the complete map-matching solution on the Zondy Vehicle Navigation System. The result show that the solution all have good accuracy real-time under the different road conditions and complex traffic, even the GPS signal is invalid or the matching road is not found. Conditional test based on criteria of distance or direction is conducted to detect the matching result automatically, make this solution has the good self-correction ability to recover from a mismatch. In all, the test is show that the complete map-matching solution this paper given can satisfy the Embedded GPS Navigation System application requirements.Finally, the paper makes a simple summary aiming at the study contents and the outlook for the research direction of the future.