Short-Term Flow Prediction of Expressway Considering Time-Space Correlation Characteristics

Short-term prediction of traffic flow is an important basis for ensuring road capacity and formulating coordinated traffic control strategies. To improve the long-term prediction efficiency and accuracy, this paper focuses on the time-space correlation characteristics of highway traffic flow based on ETC gantry and video checkpoint data and proposes a cross-sectional flow prediction model based on Long Short-Term Memory (LSTM) algorithm. The proposed model was trained and validated using actual data from a certain highway. Three evaluation indicators, root mean square error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE), were used to compare the prediction results with KNN, RF, and SVR model algorithms. The results show that the model has higher accuracy, better prediction reliability, and better comprehensive prediction performance, with the RMSE of the validated road section of 5.78, which is 22.8%, 13.5%, and 21.0% lower than the other three methods, respectively.