The Study Of Bus Dwell Time At Bus Station

As "Public transport priority" has established in China, travelling by bus is an essential option for everyone to take daily trips. Bus stations, where serve the passengers directly, are located in everywhere around the city. While the traffic flow in the city nearly reaches a saturation value recently, the interaction and influence between the process of bus dwell and traffic flow are significant, especially during the peak hour. The bus station has been the key factor which affects the traffic flow and the capacity of road and intersection. Also, the proportion of bus dwell time in travel time is getting larger. Therefore it has great significance to research bus dwell time in detail, which could help us redefine and analyze the capacity of all kinds of bus station and support us to optimize the bus dispatching and public transport network design effectively and accurately.In order to achieve these goals, this thesis did some thorough and comprehensive research in bus dwell time. On the basic of existing research, this paper firstly summarized the characteristics of bus stations classified by different definitions. Then several traditional and latest capacity models of bus stop were discussed, which made a basic foundation for setting up the bus dwell time estimation model.The progress of bus dwell was separated into three parts, entering, service and exiting, and studied in detail. Considering about the difference of curbside stop and bay stop, the microscopic behavior in bus station were emphasized, contributing to build a more correct and precise model. Also time-variant of bus dwell time was concluded by analyzing the survey data.On the basis of bus dwell time basic model, the compound Poisson bus dwell time model was proposed. Considering about the form of bus station, this model should be discussed in terms of curbside stop and bus bay. This model took the four different scenarios and the progress of exiting in bus bay into consideration in order to raise prediction and the reliability of estimation. At last this model was solved quickly by iterative algorithm.The proposed model and algorithm were validated by the data from two stations in Shenzhen. Compared with the basic model, the performance of the compound Poisson bus dwell time model was much better in describing the influence caused by queuing phenomenon and the solving progress with iterative algorithm is more quickly.