Research On The Bus Arrival Time Prediction Model

The prediction of bus arrival time is the travel service need of urban residents,and it is one of the most important traffic information for the residents,its forecasting accuracy directly affects the passenger’s satisfaction with the public transportation system.The improvement of the forecasting accuracy can save the waiting time of passengers,and can improve travel efficiency,enhance the level of public transport system services,so as to attract more passengers to take the bus,finally it can help solving the problem of traffic congestion.The bus arrival time prediction problem is studied in this paper.The arrival time of the bus is not only related to the current arrival time,but also to the historical data,based on this and according to the advantages of LSTM network model in time series analysis in deep learning,the model of bus arrival time prediction based on LSTM network is constructed,and the validity and feasibility of the model are verified by experimental method.The main work of this paper is as follows:1.The existing bus arrival time prediction models are introduced.The relevant content of deep learning and the structure of the LSTM network are analyzed in detail,the static and dynamic factors that affect the transit time of the bus are analyzed and summarized.2.A bus arrival time forecasting model based on LSTM is established.Taking the bus arrival time forecast of Qingdao city as an example,based on the processing of data deletion and normalization,a time forecasting model of bus transit based on LSTM is constructed,and the time of bus arrival of Qingdao city is forecasted,the result shows that the prediction model is feasible and effective.3.Optimization model.The model is optimized in order to improve the prediction accuracy.Finally,the data is extended by migrating learning to solve the problem of training data samples are relatively small,so as to improve the performance of the network model and improve the prediction accuracy.