Research On Freight Traffic Demand Forecasting Based On Truck GPS Data

In recent years,with the explosive growth of new forms of business such as online consumption and the improvement of residents’ requirements for logistics,the demand for freight and the volume of freight traffic has increased rapidly.it can provide a basis for accurately assessing the status quo,updating the calibration model in time,quickly realizing the forecast of freight traffic demand,and support for urban logistics planning.Aiming at the problem of freight traffic demand forecasting,this study proposed a method of constructing a freight traffic model based on the GPS data of trucks under the framework of "four stages" on the basis of sorting out the theory of freight traffic demand and applied this method to the actual planning management in Wuhan.The main research contents of this paper are as follows:(1)The characteristics of urban freight traffic demand were sorted out,and the optimization method of the four-stage model based on the GPS data of trucks was theoretically analyzed.The method of extracting truck stop points,OD pairs,and road freight flow by analyzing truck GPS data provided data processing flow for building models.(2)Based on analyzing the generation mechanism,spatial distribution,and influencing factors of freight traffic,the paper integrated truck GPS data with the DBSCAN algorithm,community discovery method,spatial multi-constraint clustering algorithm,and other new technologies and proposed the method of the freight traffic analysis zone division.(3)Based on the traditional "four-stage" freight traffic model framework,the paper compared the classic models in each stage,selected and optimized the model according to the data characteristics and theoretical basis of the GPS of trucks.Thus this study proposed the steps and methods to construct freight traffic model steps using truck GPS data.(4)Using the methods above,the paper constructed the forecast model of freight traffic demand in Wuhan.Then,the model was used to evaluate the preliminary scheme of logistics planning in Wuhan.Based on the results,the paper put forward suggestions for the planning scheme.In the logistics planning of Wuhan,this method was used to divide the freight traffic analysis zone and build the freight demand model.The results showed that: Based on the GPS data of trucks,the traffic analysis zone was divided.And the traffic area within 15% of the traffic area within the area reached about 80%.In the stage of trip generation,this study integrated multi-source data,and the model interpretability of the urban and overall freight traffic generation has reached 0.5 or more.Comparing the regression results of the scale of the traffic analysis zone with the grid-scale models,the interpretability of the traffic analysis zone-scale model is higher,reaching 0.67.In the stage of trip distribution and traffic assignment,through the simulation and comparison of the traffic assignment model of the cost function with different cost weights,the parameter calibration of the generalized cost function of the route in Wuhan was finally determined.After the current freight traffic model was established,it was applied to different scenarios of logistics planning and freight management measures to forecast the freight traffic demand in Wuhan’s planning years.Finally,the paper proposed suggestions for the planning scheme according to the forecast results of freight traffic generation,distribution,and road freight flow.