Short-Term Traffic Flow Prediction Based On Deep Learning Network

Abstract

Short-term traffic flow prediction is of importance for traffic control and guidance, and it also plays a crucial role in the development of the management and maintenance of the cross-sea great bridge. Therefore, this paper proposed a short-term traffic flow prediction method based on the data of operating private cars and minibuses on the bridge of Chang Tai expressway and a variety of LSTM network. The main work includes: cleaning the abnormal data of the original data and calculating the traffic time series in the period of five minutes, then fill in the missing data with the average of history traffic flow data. Further, the prediction model based on the LSTM algorithm is used to forecast the traffic flow of cars operating on the highway. Finally, the prediction model is tested in four different traffic conditions and the results indicate that the prediction model achieves high accuracy and generalizes well.