Study On The Prediction And Control Strategy Of Bus Operation Status Based On GPS Data

Regular bus transit is an important part of urban public transport,but also one of the means of transportation to meet the travel needs of residents.It is also an important development strategic objectives to alleviate urban traffic congestion.But nowadays,due to various uncertain factors,the phenomenon such as bus cluster frequently occurs,which leads to the increase of the average travel time of passengers.This is an urgent problem in the operation and management of conventional public transportation.In view of this situation,based on the analysis of the bus operation,this paper uses the continuous time headway to determine the status of the bus operation,which can get the characteristics of public transport in the whole process of operation.Then the paper builts the prediction model of the bus operation status and establishs the real-time feedback control system of public transport based on the operation state,so as to solve the problem of instability of public transport and improve the attractiveness of the public transport system to residents.This article first discusses the characteristics and influencing factors of the bus operation process,and explains the five common bus operation instability phenomena,and then gives the process and reasons of the abnormal state of the bus.Then the paper introduces the idea of dividing the operation state of public transportation and divides it into five operation states: bus bunching,bus bunching transition,normal state,large interval,and large interval transition;the continuous time headway is designated as the judgment index of the bus operation state.Finally,the k-means clustering algorithm is used to determine the threshold range of different states.After that,the bus GPS data that needs to be used in this article are discussed.The GPS data of public transportation vehicles contains relatively complete vehicle trajectory information due to its high transmission frequency and small positioning interval.This article first preprocesses the bus GPS data to obtain valid and usable data.The GPS data of Xi’an 700 bus is used to introduce the calculation steps of continuous time headway,and then obtains the threshold range of the bus operate state of the line through the k-means clustering algorithm.Finally,according to the continuous time headway,the variation law of headway of bus is studied.Then,on the basis of the previous research,BP neural network model,multiple logistic model and Kalman filter model are selected to build the prediction model of bus operation state,and the factors such as headway,departure interval,mileage,time period and date of the bus at the last moment are considered as the input variables of the model.The validity and accuracy of these three forecasting models are verified by the data of 700 buses in Xi’an.The final results reflect that the prediction model based on multi logistic model is better.Finally,based on the prediction results of the bus operating state,combined with the existing bus scheduling control strategy,the adjustment strategy of bus speed and delay time under different operation states is studied.After that,the real-time feedback control system for public transport based on operating status has been established to improve the stability of public transport operations.