| With the rapid development of Beidou/GPS positioning technology,vehicle networking technology,5G and other advanced technologies,today,the massive traffic trajectory data accumulated has reached the application scale of big data,which makes the traffic road network a dynamically characterized topological structure.However,the current calculation models for in-depth exploration of the evolution path of urban road network have not yet formed a complete theoretical system,and most of them remain in the study of evolution models and evolution mechanisms,and few of them can be directly applied and solve the practical problems encountered in the construction of urban roads.Therefore,this article focuses on the research on the key technologies of the evolution path of the urban road network topology,and through the self-built nodes and the data structure of the road section topology,it focuses on the construction of the urban road node and the road section classification model based on trajectory data,and then provide theoretical basis and decision support for future urban road planning.The specific research work is as follows:(1)Characteristic analysis and preprocessing of floating car trajectory data.Trajectory data is a spatio-temporal sequence that is discrete in both time and space domains under the constraints of the road network.It has the characteristics of sparseness,skew distribution,time-space sequence,nonlinearity,and non-stationary.Since low-quality data will have a serious impact on the results of mining analysis,this article has deeply studied the trajectory data acquisition system,basic attributes and characteristics,designed a trajectory data cleaning algorithm,and considered the difference of the local density of trajectory points for the detection of outliers,and proposed LOF(Local Outlier Factor)algorithm for identification and removal,laying a solid foundation for further research.(2)The construction method of node grade division model in urban road network.In order to extract the node importance information hidden in the track data of road intersection,this article innovatively proposes a method based on floating car trajectory data and based on the constructed node data structure to calculate the frequency of vehicles passing by a node at a certain time(practical significance: the importance of the corresponding intersection)and the average dwell time(practical significance: The congestion degree of the corresponding intersection)of the two evaluation indicators,and the quantitative analysis of the node is realized,then,a hierarchical model of nodes is constructed by using the Boston Matrix,which is verified by the historical image data of Google Earth.The experimental results show that the accuracy of the model is about 85.71%,the reliability and validity of the model are further illustrated.(3)The construction method of link grade division model based on node topological connection relation in urban road network.In order to mine the road section importance information contained in the track data on the road,this paper innovatively proposes a method to start with the track data source,by calculating the traffic frequency(practical significance: the importance of the road section)and the average passing speed(practical significance: The congestion degree of the road section)in the road network topology,the evolution process of the road segment in the actual road network is reconstructed,the quantitative analysis of the road segment is realized,the road segment classification model is constructed by using the Boston Matrix,and is verified by the historical image data of Google Earth,the experimental results show that the accuracy of the model is about 72.73%,which proves the reliability and validity of the model.In the end,the experimental results can be fed back to the construction,planning and management of urban roads,so as to further achieve the purpose of guiding the evolution of the transportation network in the direction of better function and structure. |
PDF Full Text Request