Application Of Kalman Filtering In Map-matching

Map-matching, refers to obtaining the bias information of carrier position through matching algorithm and real-time revising, demonstrating accurately position of vehicles in road net. Map-matching algorithm, affecting on the precision of vehicles location, is a key technology of directly deciding final performance of navigation products.At present, many scholars proposed different map-matching algorithms, respectively have good and bad points in reducing systematic positioning error aspect, but mostly involves quite few to kind of errors of instrument survey data, electronic map data and coordinates transforming, which play a highlights part in insufficiency and bad effects on final matching quality. The vehicles navigation localization belongs to a kind of dynamic system, bringing out with series of data along time passage, therefore, seeking one method to improve processing result to the matched data dynamic error, heighten matching quality has the necessity very much. The dynamic filtering technology is a data processing method with widespread application foundation in the engineering, introducing the Kalman filtering into map-matching process, for its superiority on the aspect of error, to improve reliability and flow of map-matching, has a high researching value.This article unifies the principle, method and characteristic of Kalman filtering (conventional filtering, decomposing filtering, adaptive filtering, nonlinear filtering and so on), inquired about way and method of applying filtering technology in the map-matching, discuss method of establishing filter suitable to vehicles mobile characteristic. Using MATLAB tool software, realizing programming map-matching based on each concrete filtering and examining application effect of the new match method by simulation experiment.With SDK tools, such as Visual C++ 6.0, MATCOM, Access and so on, an map matching visualization software is developed, realizing map matching algorithm based on Kalman filtering designed, at the same time, providing a software platform for the comparing between different map matching algorithms, and also building the good foundation from now on for further the thorough research map matching algorithms. Through the experiment, the comparison of new matching algorithm which designed in this article with traditional map matching algorithms (such as linear projection), confirm the validity of this algorithm.