Research On The Bus Arrival Time Prediction Based On GPS Data

With the continuous advancement of urbanization,car ownership and residents’ travel demand has surged,which has caused the problem of urban traffic congestion to become more and more serious.Vigorously developing public transportation is an effective and economical way to solve urban traffic congestion.Improving public transport service level is one of the important measures to implement public transport priority.It is particularly important for bus passengers to provide accurate and reliable information about bus arrival time.However,the current in-depth research on bus arrival time is relatively scarce.This article divides bus arrival time into two parts: stop time and travel time between stations.Based on the analysis of the two parts,this article builds a prediction model of bus arrival time.Firstly,this article introduces the GPS data collection process and error analysis,and designs an algorithm for matching GPS data with bus stops.Next,this article analyzes the operating characteristics of public transportation vehicles.The article also analyzes the driving laws of public transportation vehicles from the perspectives of driving time,the same period on the same day,and the historical period.These work laid the foundation for the construction of the prediction model of bus arrival time.Secondly,this article analyzes the stopping time and travel time between bus stations.Based on the analysis of the factors affecting the bus stop,this article proposes a station stop time prediction model based on the KNN algorithm.In this article,three methods of BP neural network,Kalman filter and combined model are used to predict the travel time between stations.The prediction results are analyzed from two aspects of absolute error and relative error,and the results show that the combined model has the best prediction effect.Finally,the article establishes a model for predicting bus arrival time.The KNN algorithm is used for station stop time prediction,and the combined model is used for station travel time prediction.Next,this article selects 700 roads in Xi’an for example verification.The results show that the prediction accuracy of the arrival time within 6 stations downstream from the current station is high,and the prediction error is within 1 minute.Considering the frequency of bus departures,the accuracy of the prediction model proposed in this paper can meet the needs of passengers to query bus arrival time information.Research results such as the analysis of the running rules of the bus and the combined forecasting model of inter-station travel time can provide some reference for the research of scholars in this field.