Research On Map Matching Algorithm In Vehicle Navigation System

Intelligent Transportation Systems (ITS), which integrate sensors, electronics, computers, data communications, Geographic Information System (GIS) and other technologies, plays a more and more important role in our lives. As the most important application system of ITS, Vehicle Navigation System (VNS) are the core of ITS. It was created to navigate vehicle especially on urban area. Due to the positioning data acquired from Global Positioning System(GPS) and Dead Reckoning(DR) suffers the measurement error, the efficiency of ITS would be reduced. Fortunately, this can be resolved by Map Matching (MM) most cost-effectively.MM, which realize an optimal estimation of the vehicle position by integrating the positioning data with the spatial road network data, is supposed to the key of VNS.Existing MM algorithms have weaknesses which decrease their capability to support VNS. Therefore, a robust and reliable MM algorithm is developed in this research.This thesis has discussed items as follow:First, this thesis discusses the importance of MM in VNS.Second, the performances of existing MM algorithms are introduced.Third, a variety of factors influenced MM are analyzed.Next, the functional modules of MM algorithm are divided.Then, the elements can be used in MM algorithm are excavated.Based on above procedures, a topological MM algorithm base on weighting is developed. In this algorithm, a simplified ellipse errors region is constructed to reduce the number of segments to match, and within the errors region, calculate each segments’weighting score, which is the summary of matching score on heading, proximity and relative position, then select the segment of highest weighting score as the correct segment, on which the optimal position is estimated.Moreover, the testing result on the road of Kunming urban area suggests that the algorithm can correct the bias of position information even at a junction.