Research On Short-time Passenger Flow Forecasting Model Of Bus Stop

With the rapid development of social economy and motor vehicles increasing,the traffic congestion has become more and more terrible.Developing the public transportation in advanced has been treated as a kind of strategy in every country which can solve the problem effectively,and they’re trying to establish a intelligent transportation system as well.Although one of the most important parts of the system is scheduling management,the forecast of passenger flowing is the basis of dynamic scheduling.The forecast of short time passenger flow at public transportation stations can reach the goal of obtaining a more accurate information of passenger-flowing forecast in the line network.Therefore,how to select the appropriate models and the forecast for short-passenger bus stop would be the critical part of this article.Firstly,this paper researches models about the forecast of Bus Stops short-term passenger flow.Then according to the comparison and analysis of different models,sum up the advantages and disadvantages of each model and the scope of each model.It is the theoretical foundation of this study.The analysis of different types short-time passenger flow data and the features of space distribution of bus stops and the time correlation and spatial correlation of bus stops short passenger flow’s features can prove that we can establish a unified prediction model of different days,and provide the basic selection of construct model input layer data.Secondly,this article concentrate on how to establish the model which bases on the improved RBF neural network predicting model to predict bus stops short passenger flow.Through the analysis of passenger flow data,We choose the relevant data as the input data for picking the model of input layer data,therefore we can reduce the input dimension.Classification is made for the input data,and the traditional RBF neural network model is established separately for different types of data at the same time.And then extract the characteristics from every type of data,and finally,we use the output value times the weight of each model sum as the final result,so that we can improve the accuracy of the forecast.Finally,the model is verified by the actual passenger-flowing data of some bus stations in Dalian,then we put forward a comprehensive forecast model for the short time passenger flow in the public transportation station,and realized the functions.Based on the above,research-based B/S architecture Short-time passenger flow forecasting application system,prove the practicality of this paper content.The experiment turns out that the model proposed in this paper predicts the short time passenger flow of Dalian public transportation station basically.The result of research on this paper has theory and practical application value to the short time passenger flow forecast of public transportation station.