GPS Data-Based Bus Corridor Travel Time Reliability Assessment Model Research

In order to objectively grasp the resource effect and service level of the bus corridor,this dissertation relies on conventional bus vehicle GPS data and establishes a bus corridor journey time reliability model with journey time as the core,and validates the proposed journey time reliability model,based on which the journey time reliability of the bus corridor is estimated and the relatively weak sections of its operation services are identified,for the sake of bus operation The proposed journey time reliability model is validated and the proposed journey time reliability model is validated.This dissertation analyses the composition structure of conventional bus GPS data,screens out erroneous data and extracts valid data,including latitude and longitude coordinates,collection time,vehicle ID and other information;with the help of sampling interval kernel density probability distribution chart,the sampling interval of the data is counted and the feasibility and reliability of the data is verified;on this basis,a map matching algorithm based on the KD-Tree nearest neighbour query and a journey time extraction method are proposed.An empirical analysis is employed to describe the volatility of travel time from both time and space dimensions,and four common probability distribution models of travel time-Weibull distribution,log-normal distribution,normal distribution,and gamma distribution-are curve fitted and parameters estimated based on GPS data.To begin,the influencing factors of travel time are examined.A log-normal distribution-based probability distribution model of travel time is established.The meaning of bus journey time reliability is firstly described,and probability-based indicators,mathematical and statistical indicators and TTI indicators are used to comprehensively assess the journey time reliability of bus priority corridors.The analysis focuses on probabilistic indicators to assess journey time reliability,combining journey time thresholds with service levels,and using the quantile method to compare the slopes of cumulative percentile plots for unit length journey times,selecting the unit distance journey time thresholds corresponding to 20%,65%,85% and 90% quantile points with large sudden changes in slope as the boundary values for the classification of service levels A,B,C,D and E.A lognormal distribution-based road section travel time reliability assessment model is thus created,which serves as the foundation for a path travel time reliability assessment model that takes into account the correlation between roads.The reliability assessment model was tested based on the empirical cumulative distribution of statistical travel times from GPS data,and the accuracy of the model was verified.A comprehensive assessment of the journey time reliability of bus priority corridor sections was carried out using probabilistic,mathematical statistical and TTI indicators based on GPS data from 84 conventional bus vehicles,and weak sections of bus priority corridor operations were identified.