Abstract
To address the problems of easy volatility, empty matching, incorrect matching, and low matching accuracy, an adaptive projection map matching algorithm is proposed based on existing map matching algorithms for intersecting urban roads. First, the method of setting the distance threshold was used to eliminate abnormal location points. Then, by comparison with the AutoNavi map, the missing data in the OpenStreetMap were filled. To reduce the matching time, the grid index was generated, and the impossible methods were filtered by the error circle. Second, the probabilities of the projection distance and direction were assigned, and the corresponding weight coefficients were adjusted adaptively. Finally, the probabilities of the candidate methods were calculated by considering both the projection distance and direction, which helped to determine the actual road and improve the matching accuracy. At the same time, real-time driving data from Changzhou taxi vehicles were used for experimental verification. The results show that the adaptive projection map matching algorithm can increase the matching accuracy by approximately 4% compared with other existing matching algorithms and shorten the single-point matching time by approximately 1.5 ms, which realizes accurate map matching under complex intersecting methods.