Research On The Predicting Method Of Bus Travel Time

Research on the predicting method of bus travel time mainly included two aspects which are travel time of stop to stop predicting model and bus dwell time predicting model. Considering our country's actual situation, this thesis has developed a bus travel time predicting method and conducted a case study through the analysis of bus operation characteristics and regularity. All of the work will improve service level of public transit system, increase public transit's attraction and optimize the structure of trip structure.Firstly, this thesis reviewed domestic and foreign literatures and summarized the existing methods for bus travel time predicting. According to the characteristic of each method, this thesis did classification analysis on them.Secondly, in order to meet the data demand of travel time predicting, a method for processing the GPS data was presented according to the characteristic of GPS data. And a computer program was developed using C#. NET and SQL Server 2000. This computer program can improve GPS data processing efficiency by converting the GPS raw data into the information like travel time, dwell time, stop times, speed and so on.Thirdly, based on the investigation of bus dwell time, the thesis conducted an influence factor analysis for bus dwell time and established a bus dwell time predicting model. The model can predict the dwell time by boarding and alighting passenger flow data and crowding degree of the bus.Fourthly, based on the result of GPS data processing, this thesis analyzed the characteristic of bus travel time and used time series theory and Kalman Filter model to establish two kind of bus travel time predicting model. Furthmore, this thesis compared these precision of the two kind of model.Finally, this thesis analyzed a case study with a real bus line in Beijing, and the results of the predicting showed that this method can complete an accurate and reliable bus travel time predicting.