Temporal And Spatial Characteristics Of Urban Traffic Based On Taxi Trajectories

At present,it is difficult to meet the needs of fast-growing vehicles by expanding the scale of urban roads.GPS trajectory data provides a new way and method to alleviate the pressure of urban traffic.In the case that the traffic pressure is more and more difficult to overcome by increasing the scale of roads,from the perspective of traffic big data,using the GPS data of floating vehicles to assist intelligent traffic is an important means to alleviate the traffic conditions of urban traffic.The method of using the floating vehicle GPS trajectory data to assist in improving the urban traffic situation can generally be started from two aspects.First,make full use of the characteristics of the road network structure,and play the role of road network traffic current.The second is to explore the spatial and temporal characteristics of urban traffic and rationally divert it to alleviate urban traffic pressure.For the former,studies have shown that increasing road network density has little effect on traffic improvement,and the research on the relationship between road network structure and traffic conditions needs further study.For the latter,the spatial and temporal characteristics of urban traffic are an important part of urban traffic and are of great significance for improving urban traffic conditions.Therefore,this paper takes Chengdu as the research area,and takes the GPS trajectory of more than 14,000 taxis in August 2014 as the data source.Firstly,the Cosine kernel density road network extraction algorithm which introduces the directional weight of the trajectory is proposed,which is used for road network extraction and complex network construction.Then,three dynamic traffic centralities were studied to analyze the urban static road network structure and the urban dynamic traffic network structure.Finally,the spatial analysis method and the space-time cube model are used to analyze the trend and mode of urban taxi hotspots in Chengdu.The main work and research results are as follows:(1)An algorithm for extracting Cosine kernel density road network introducing GPS track direction is proposed.The Cosine kernel density algorithm takes the network kernel density algorithm as the idea,and introduces the directional weight of the trajectory in the kernel function to extract the connectivity properties of the road network.and subdivides the region to improve the extraction accuracy of the low-grade road.The cosine kernel density road network extraction algorithm is used to extract the internal network of Chengdu Sanhuan Road.The results show that the accuracy of the extracted road network reaches 83.2%,which can effectively extract the connectivity between road nodes and road segments,and improve the traditional nuclear density algorithm.The shortcomings of the road details.(2)Three dynamic traffic centralities that take into account traffic flow characteristics,road geometry and road connectivity are proposed.Traditional central indicators can effectively analyze the static urban road network structure,but it is difficult to evaluate the dynamic urban traffic network characteristics.This paper proposes three dynamic centralities that take into account traffic flow characteristics,road geometric features and road connectivity.Sex indicators,dynamic traffic centrality can effectively analyze the importance of local traffic nodes,dynamic traffic proximity and intermediary centrality can effectively analyze the importance of global traffic nodes.(3)Using the theory and method of complex network,analyze the topological structure characteristics of Chengdu urban road network and the structural characteristics of urban transportation network.The characteristics of road grid topology are analyzed by degree centrality,near centrality and intermediate centrality.The results show that Chengdu road network is not a scale-free network.The center of the network is the center of the geographical location,and there are not many nodes acting as "bridges".Only a few nodes have a greater control over the entire road network;The paper analyzes the structural characteristics of urban road traffic network with three dynamic traffic centralities,and finds that Chengdu traffic network has the characteristics of scale-free network.There is a serious heterogeneity between nodes,and only a few nodes undertake important traffic tasks.For random traffic accidents or traffic congestion,there is a good confrontation.The center of the traffic network becomes the loop and the central city.The node acting as the;"bridge" is located on the loop,and the traffic flow is well-distributed.Traffic pressure on the downtown area.(4)Analysis of hotspots,traffic hotspots,hotspots and hotspots of Chengdu taxis by spatial analysis and space-time cube model.We first use the method of statistical analysis to study the overall law of taxi trips.Then based on the space-time cube model and spatial analysis method,the spatial distribution and space-time hotspot patterns of the passengers and passengers are analyzed.It is found that the taxi city traffic has weekly and daily cycle effects,and the spatial distribution hotspots of traffic flow can be divided into three types,while the hotspots of the passengers are concentrated in four types of areas,and the hotspots and traffic hotspots at diff-erent ti,mes.There are big differences in trend changes.Analysis of hotspot patterns and changes can better facilitate taxi services and residents' travel and improve road traffic pressure.