Analysis And Prediction Of Traffic Status In Chengdu Based On Taxi GPS Data

The research on urban traffic condition has always been a key point in urban problems,and how to guarantee the smoothness and safety of urban traffic is its core.At present,under the condition of well-equipped intelligent traffic hardware facilities such as positioning and navigation technology,digital communication technology,electronic sensor technology,etc.,24-hour continuous monitoring of road condition and vehicle condition has been realized.At present,the real-time monitoring data of the road network can be processed and processed and released to the navigation platform to provide drivers with real-time road conditions,guide the travel paths of traffic participants,and improve the capacity of the road network.However,on the other hand,if the real-time monitoring data are deeply analyzed to correctly predict the traffic flow changes of the road network,the traffic management department can take the initiative to release information,broadcast,control signal lights,artificial traffic control and other ways to ensure the smooth flow of traffic and change passive adaptation into active regulation.Now,the technology base and data base for the further research of the urban traffic provides support,so this article to Chengdu in August 2014,a taxi GPS data within and around the city of Chengdu high-speed road network as the experimental data,through a series of data processing and analysis of traffic flow state of Chengdu information and characteristics of time and space,and through the ARIMA model and STARIMA model to predict traffic flow parameters of the whole network.The specific contents can be divided into the following aspects:(1)Data preprocessing and calculation of traffic flow parameters.Through GPS data processing,map matching,the establishment of Chengdu road network model,traffic flow parameter selection and calculation,the traffic flow state information statistics of each section of Chengdu city road network is realized.(2)Analysis of space-time characteristics of traffic flow in Chengdu.According to the characteristics of road traffic flow and the topological structure of road,the spatial weight matrix of road network traffic is constructed,and the characteristics of traffic flow in Chengdu are explored from the perspective of time series characteristics,spatial distribution characteristics and spatial-temporal correlation analysis.(3)Traffic flow forecast of Chengdu ARIMA model and STARIMA model were constructed according to the four basic steps of sequence stabilization,pattern recognition,parameter estimation and diagnostic verification,and the traffic flow parameters were predicted.Finally,the prediction effect and performance of the two models were compared and the conclusions were drawn.The results of this paper show that the clustering analysis based on time series can obtain two kinds of traffic patterns in time dimension and four kinds of different road types in space dimension.The road network spatial weight matrix based on traffic flow influence rule performs better than the spatial weight matrix based on Rook rule in spatial correlation analysis.The traffic flow in Chengdu shows the characteristics of spatial and temporal correlation in different directions and positions of the loop road and the longitudinal road.