Passenger Flow Prediction of Shenzhen Metro based on ARMA Model

Abstract

Due to the speeding up of urbanization and the diversification of urban transport modes, the demand for urban transport in people's lives is also increasing. In order to alleviate the pressure of urban traffic, cities are vigorously developing public transport, especially metro transport. This paper will use the metro card swipe data of the evening peak on 1 September 2018 in Shengzhen, combined with the video recognition technology, to optimize the ARMA-based model to forecast the metro's short-term passenger flow. Based on the above data and methods, this paper takes the change volume of inbound passenger flow from 17:00 to 20:00 as a reference, and conducts research on the metro short-time cross-section passenger flow prediction. It is found that the ARMA-based short-time passenger flow prediction model can provide timely and effective feedback on the changes of patron movement in the waiting area. The short-time cross-section customer flow prediction of metro helps to alleviate the congestion of the peak, improve people's living standard, save the cost of travelling time, and is of practical significance to the urban rail transport operating companies.