| With car ownership has continually increased,urban traffic congestion has become an urgent problem in today’s society.Floating car technology has become one of the main means of urban road traffic congestion monitoring,because of its wide collection area,less installation costs and relatively stable collection system.However,floating car data is limited by its own technology,it is susceptible to positional distortion,drift,and other factors,and data processing which is called map matching is required to restore data authenticity.The frequent congestion has an inestimable impact on the environment,economy and people’s livelihood of the society.How to alleviate the frequent occurrence of congestion has become an urgent problem to be solved in today’s society.Traffic state research can effectively mine the laws of urban traffic conditions and provide powerful theoretical support for traffic managers to formulate mitigation systems and policies.The basis of traffic state research is traffic detection data.Currently commonly used traffic detection data includes: fixed detector data and motion detector data.The fixed detector data is obtained by the road detection equipment installed on the road facilities.The initial investment and maintenance cost of the fixed detector are too high due to the problems of domestic road construction and high reconstruction frequency;the floating vehicle technology is installed by the GPS device on the vehicle collects data and has the advantages of wide coverage area,low installation and installation cost,and relatively stable acquisition system.It has become the main data source for urban road traffic condition monitoring.However,the floating car data needs to establish constraints according to its own positioning data and the correlation of multi-attribute road network data such as roads and intersections,and find the travel path that conforms to the road network structure,that is,the map matching process.In the existing map matching research,there is a problem that the data of the massive floating car cannot balance the matching accuracy and matching efficiency.In this study,a map matching algorithm based on quadratic grid and feature weighting is proposed.The secondary grid efficiently selects candidate segments to determine the shortest path.The feature weighting takes distance as the main factor,and the driving direction and trajectory angle are adjusted by speed and interval length.The secondary element accurately determines the optimal matching road segment to overcome the problems of parallel road segment jump,intersection mismatch,dense road segment mismatch and large fault tolerance when the positioning error is large,and to avoid the driving direction and the trajectory angle failing at low speed.The interference caused by the map matching,the driving direction and the trajectory angle element dynamically adjust the weight by the instantaneous speed and the interval length between the trajectory points.The experimental results show that the improved feature weighting method in the proposed algorithm significantly improves the matching accuracy.In the algorithm principle,the quadratic raster method is combined to effectively improve the algorithm matching efficiency.The map matching algorithm proposed in this study can face massive data.Match the matching accuracy and matching efficiency.The traffic operation state data after the floating car data processing—the segment speed data is time series data,the data dimension(time dimension)is high,and the clustering algorithm has better data analysis performance when facing high dimensional data.In the study of urban traffic operation state,the clustering algorithm is used to analyze the traffic operation state data and explore the potential law of urban road traffic state.However,in the existing research,there is a problem that the algorithm has poor anti-noise,ignores the potential law of urban road traffic state under the spatial dimension,and lacks research on the main road.In order to overcome the potential law under the space-time dimension of urban road traffic and overcome the shortcomings in previous research,this paper proposes a K-Means clustering algorithm based on wavelet denoising to overcome the problem of poor anti-noise performance of the algorithm;Cluster analysis,to explore the difference between the clustering results between the various segments,the whole and the composed segments,cluster analysis of the bidirectional velocity trend of the main road speed in the time dimension,compare the results,in the time and space dimensions Digging the operational characteristics of urban roads to make up for the shortcomings of previous studies. |
