Research On The Arrival Time Prediction Method Of Modern Streetcar Based On The LSTM Model

In order to alleviate the pressure of urban traffic congestion and reduce t traffic pollution,China drawing on the experience of foreign large cities in traffic development,attaches great importance to the construction of large volume public transport system,in which the semiindependent right of way modern streetcar having the advantages of low construction cost,small construction difficulty,safety and environmental protection,strong flexibility,etc.,has been rapidly developed in China in recent years.The semi-independent right of way modern streetcar has many advantages,but it is easy to be affected by delays such as signal control at intersections along the line,which leads to the reduction of the reliability of the streetcar travel time and the difficulty for passengers to obtain the accurate arrival time of the streetcar.The arrival time of public transport is the most concerned information for the most passengers,and the reliable arrival information has a statistically significant inhibitory effect on the perception of waiting time of passengers.Therefore,this paper aiming to improve the accuracy of tram arrival time prediction,enhance the competitiveness of streetcar,and increase the ratio of public transport trips,carried out the research on the streetcar arrival time prediction.This paper took the semi-independent right of way modern tram as the research object and time series data as the point of penetration,proposed one arrival time prediction system which considering the spatiotemporal correlation in the operation process of modern streetcar.The system can identify the complex spatiotemporal patterns that cannot be captured by traditional prediction methods.The core of the system is a deep neural network,which is composed of the long short-term memory neural network and convolutional neural network.This paper made an empirical evaluation of the designed travel time prediction model of streetcars and compared it with other mainstream prediction models.The results show that the model has high accuracy and reliability in predicting the arrival time and travel time of streetcars,especially in complex traffic conditions(such as peak hours),the model also has excellent prediction performance.