Study On Prediction Method Of Arrival Time Of Urban Public Transport Vehicles

According to the national economy's development,China's urban population's increasing,and the per capita car ownership' increasing,the urban motor vehicle ownership is continuing to grow.However,the speed and scale of the existing transport infrastructure can not meet the needs of the growing urban transport.Meanwhile,the contradiction between supply and demand makes the city traffic increasingly congested.Therefore,in order to solve the urban traffic problem,it need to develop urban public transport.At this stage,conventional buses are a major component of public transport in many small and medium-sized cities.Conventional bus is the most extensive coverage,the largest number of running lines,and the lowest cost of travel to solve a city residents' best way to travel.However,conventional bus travel ratio is not high and the bus attractive is low for the residents in many small and medium-sized cities.The main reason is the regular bus' arrival time is uncertain,punctuality is poor,passengers need to wait for unknown time,need to pay attention to the station of vehicle information,and bus deliver information is more backward,appearing "extension of the neck" and other public transport status.Passengers prone to anxious in the processing of waiting bus,or directly change the other communication media.Therefore,in small and medium-sized cities,accurate forecasting the bus arrival time can improve the proportion of residents traveling,improve the satisfaction of passengers and improve the service level of urban public transport,which is of great significance to solve the traffic problem.In this paper,the operation characteristics and influencing factors of bus arrival time are analyzed,and the arrival time of urban bus is divided into the time of travel,the time of stopping and the delay of intersection,three parts.The factors of bus arrival time are selected,based on the analysis of the operating characteristics and influencing factors of these three parts.According to the GPS data generated by the bus operation,the algorithm of data processing and the algorithm of bus linear matching are designed.Then,the paper analyzes the theory of bus arrival time prediction which is commonly used at home and abroad.In this paper,the first step is on the basis of analyzing and gaining a certain amount of bus arrival time data,using of bus vehicle GPS' historical information and real-time information to deal with the model,and then get the bus station's history arrival time-table,the real-time of running speed and bus vehicle location and other information.The second step is on the basis of analyzing the bus vehicle operating characteristics,establishing the busvehicle's time model on the road running time,site docking time,and intersection's delay time.The third step is to use the Kalman filter's high accuracy in the short-term prediction,predicting the bus next station arrival time.Based on the previous preparation,it establishes the prediction model of bus vehicle GPS information of arrival time-table.Finally,the bus vehicle arrival at the previous station' real-time and predicting time is as the input of Kalman filter,and use Kalman filer recursive equation to predict the bus vehicle arrival time.On the basis of Kalman filter's dynamic real-time arrival time-table predicting model.In order to validate the proposed model,this paper selects three representative bus lines in Nanchang.Through the acquisition of Nanchang 51 Road,220 Road,521 bus running 11 days of GPS data,and the car follow data in three lines of each site actual depart and arrival time,the GPS data is processed according to the proposed model,then gaining the bus vehicle's running time,speed and location.The paper uses the examples to verify the establishment bus vehicle' arrival time predicting model,and uses the average absolute error and the average absolute error percentage as the index of error analysis.Through the comparison and analysis of the model predictions and the actual values,using the error evaluation index's evaluating results,it can be concluded that the bus arrival time forecast accuracy are more than 90% whether on the peak or non-peak in the three different characteristics of the bus line.Therefore,the constructed forecasting model in this paper has high prediction accuracy and applicability,the model operation is more simple and the real-time effect is better,which can be used to forecast the arrival time of city bus.In this paper,the process of information dissemination is discussed,which is released by multi-channel to passengers,so that the research results can realize the application value better.