Improved Map Matching Algorithm Based On Dynamic K Nearest Neighbor And Historical Matching Data

The Global Positioning System(GPS)is a high-precision radio navigation and positioning system based on aerial satellites and ground base stations,and is an important way for electronic devices to obtain geographic locations.However,affected by the accuracy of the equipment and the external environment,there is a deviation between the GPS positioning data and the actual position,which affects the application effectiveness of the Intelligent Transportation System(ITS).In order to solve this problem,a map matching algorithm projected on the matching road segment after correction of GPS positioning data is an effective way.The core of the existing map matching algorithm is to analyze the correlation between the GPS positioning data with time series and urban road network data to find the correlation between the two to achieve matching,when the structure of the road network where the matching point is located is more complicated,it will be misjudged.In response to the above problems,the data features contained in historical matching data were used to enhance the performance of the algorithm.And an improved map matching algorithm based on dynamic k-nearest neighbor and historical matching data was proposed in this paper.The main work of this article is as follows:(1)The experimental data used in this paper is derived from the GPS positioning data of taxis provided by the Xi'an Transportation Bureau.This data is the original data,which was preprocessed,and the historical matching data was associated with the matching road segments.(2)A dynamic k-nearest neighbor error offset correction algorithm was proposed.The algorithm relies on the principle that the vector features between the positioning data in adjacent areas have similarity,and the data was directionally shifted toward the matching road section,thereby improving the matching accuracy of the positioning data.The algorithm consists of three stages: data normalization processing,dynamic k-value network model training and data error offset correction.Compared with the existing algorithm,it can be known that the algorithm can better select the similar points of the points to be matched,and exhibit a better error offset correction effect.(3)A geometric map matching algorithm based on historical matching data was proposed.The algorithm performs further matching on the basis of data error offset correction.It mainly includes the following processes: first,the improved meshing algorithm was used to determine the set of candidate road segments for the positioning point.Second,the relationship between the positioning point and the historical matching data was used to determine the set of candidate road segment.Finally,the weighted sum of matching similarity,projection distance,heading angle,trajectory angle and shortest path distance was used to obtain matching road segments.Experimental comparison with the existing algorithm shows that the accuracy of the algorithm is improved compared with similar algorithms.