Research On Low-Sample-Rate Map-Matching Algorithms For Vehicular Ad Hoc Networks

In vehicular ad hoc networks,many applications need to know the position of vehicle.But the positioning system is influenced by any factors such as the weather and city infrastructure.As a result,positioning errors in the system are usually happened,even lose the local information,which have serious impact on the application performance.Map matching is a technology which can make use of the digital map to adjust the vehicle location in road network.Map matching provides a new way to solve the problem caused by inaccurate positioning in vehicular ad hoc networks.This thesis studies how to reduce the impact of vehicle position errors from positioning system with map matching techniques,in order to provide the accurate location position to the applications of vehicular ad hoc networks.The main works of this thesis are as follows.The importance of map matching in vehicular ad hoc networks is explained and studying the status quo of map matching to analyzes the principle and evaluation method of map matching technology.The existing high-sample-rate map matching algorithms are discussed form the low-sample-rate datasets objective point of view,and the shortcomings of these algorithms are proposed.A map matching algorithm(MIV-Matching)to eliminate miscalculation based on low-sample-rate data is proposed.MIV-Matching analyzed the reason of miscalculations in existing low-sample-rate map matching algorithms,and given a method to avoid miscalculations,which can improve the accuracy of matching results.By using real GPS trajectory data to evaluate the MIV-Matching algorithm,the results showed that the error of MIV-Matching is reduced 50%than existing low-sample-rate map matching algorithms.A low-sample-rate map matching algorithm(Dir-Matching)that considers the influences of vehicle direction is proposed.By researching the process of deal with the GPS trajectory datasets of existing low-sample-rate algorithms,find the matching operation failed to fully use of the datasets information.On this basis,Dir-Matching considers the geometry-topology analysis process of existing low-sample-rate map matching algorithms,and combined with the connection between direction of the vehicles moving and roads.The experiments show that the Dir-Matching can significantly reduce the errors and improve the accuracy of matching results.An effective method of programing by Java and using Google Earth to evaluate the accuracy of map matching algorithms is proposed.