| Bus stops are important nodes of the bus network.Service time and queuing delay time of buses at the bus stop play an vital role in evaluating the service reliability of buses and estimating the capacity of the station,and they are also key indicators to help managers in analysis and decision-making.AVL system has provided massive GPS data for bus operation and management,but the sampling frequency of GPS data in most cities is low,which cannot accurately capture the movement details of buses while serving the bus stop such as stopping behaviors.Therefore,based on the probabilistic model proposed by Hao,this paper established a trajectory reconstruction method suitable for the process of buses entering and leaving the bus stop to estimate the second-by-second location,speed and motion state of buses during the process by analyzing the operation characteristics of buses.The specific research is as follows:(1)The method of collecting and processing the complete trajectory data of buses entering and leaving bus stop for parameter calibration and model verification is introduced.The video data of buses entering and leaving the station is collected by high-point shooting,and imported the data into the Data From Sky Viewer software for analysis to obtain the complete trajectory data in pixel coordinates.Then coordinate transformation,abnormal data processing,trajectory smoothing processing and motion state recognition are carried out.At the same time,the basic information and processing flow of bus GPS data are introduced.(2)Based on the collected complete trajectory data,the qualitative and quantitative analysis of the operation characteristics of buses entering and leaving the station is carried out,and the change rule of the motion states during buses entering and leaving the station is obtained.On this basis,by assuming the distribution of key variables during bus operation,a probability estimation method is proposed to estimate the probabilities of all the possible trajectories,and the trajectory with the highest probability is used as the final reconstruction trajectory.(3)Parameter calibration and model verification are performed based on the collected complete trajectory data.The model is tested from the two perspectives of trajectory reconstruction results and stopping behavior estimation results,and a comparative analysis is made with the reconstruction results obtained by cubic polynomial interpolation and the original probability method proposed by Hao.The results show that the adjusted probabilistic method can more accurately capture the stopping behavior of buses,and the estimation of the stopping start time,location and stopping duration is also more accurate,and can better reflect the real process of buses entering and leaving the bus stop.The applicable boundary of the model is also discussed.(4)Using the real GPS data as the input of the model,the complete trajectory data of the bus entering and leaving the station is obtained.Then,based on the spatiotemporal trajectory relationship of the front and rear buses,a bus stopping type judgment method is proposed for the single-berth stop and the multi-berth stop respectively.The stopping types mainly include queuing,stops and other.Based on the judgment results,the queuing delay time and stopping time of the buses are further calculated.At the same time,based on the complete trajectory data,the headway of each line at the corresponding station is calculated and analyzed. |
