Research On Route Choice Behavior Of Urban Freight Vehicles Based On GPS Data

The phenomenon of the emergence of a large number of logistics vehicles brought by urban productive service activities,along with the rapid development of new logistics modes such as crosstown freight,express delivery,and network freight platforms,has led to freight vehicles becoming an important part of the urban transportation system,and the rapid growth of the number of freight vehicles has exacerbated the traffic pressure in cities.The travel behavior of urban freight vehicles is different from other types of vehicles,and it is of great practical significance to reveal the characteristics of urban freight vehicle travel behavior and route choice mechanism based on large-scale trajectory data to solve the practical problem of lacking data mining technology and route choice model support for the planning and operation control of urban freight transportation system.This thesis deeply investigates the route choice behavior of urban freight vehicles based on GPS data,involving three sub-problems of GPS data pre-processing,route choice set construction and route choice behavior modeling,and the main research work and conclusions are summarized as follows.1)GPS data pre-processing.The combination of Python programming and Arcgis software is used to preprocess the travel data and combine with implicit Markov map matching algorithm to obtain travel information such as travel time,travel route length and route detour in urban freight vehicle travel.2)Alternative route choice set construction.The DBSCAN algorithm is used to spatially cluster the origin and destination points of truck trips in Changchun,and generate the initial route choice set,covering 27 truck travel hotspot OD pairs with a total of 7713 travel data;the improved BFS-LE algorithm is designed to add the route search rules and construct the alternative route choice set.The algorithm improves the operation efficiency of the algorithm and the quality of the generated route choice set.3)Route choice behavior modeling.Firstly,based on the driving information of freight vehicle travel,the spatial and temporal characteristics of urban freight vehicle travel were mined and analyzed,and the regular characteristics of urban freight vehicle travel and factors that may affect urban truck route choice were obtained.Secondly,the MNL and PSL route choice models were constructed respectively,the main parameters of the models were calibrated,and the effectiveness and prediction accuracy of the models were verified by combining the goodness-of-fit,significance analysis results and hit rate of the models.It is found that when truck drivers choose travel routes in urban roads,the travel time,detour degree,the number of left turns and intersections,and the proportion of major and minor roads in the route have negative utility effects,and the number of right turns,the proportion of feeder roads and the route correction term in the route have positive utility effects.both the MNL model and PSL model can better explain the route choice behavior of urban freight vehicles,and the PSL model is better than the MNL model in explaining the route choice behavior of vehicles.This thesis provides a model reference for the data-driven urban freight vehicle route choice decision based on the urban freight vehicle travel pattern revealed by GPS data,the construction of an effective route choice set and the route choice model.