Research on Short-Term Traffic Flow Prediction Method on Expressway

The short-term traffic flow prediction of expressway is of great significance to the intelligent management and control of expressway. By summarizing the research contents on expressway short-term traffic flow prediction in different literature, the shortcomings of the current short-term traffic flow prediction research are found; the short-term traffic flow prediction process is given. Then the short-term traffic flow prediction models are classified and compared. The applicable scenarios, advantages, and disadvantages of different models are clarified, and the processing methods of missing data and wrong data are studied. Through the analysis of specific case data, the prediction accuracy of KNN model, SVM model, and LSTM model is compared. It is found that the prediction accuracy of KNN model is the highest, and the MAPE value is 29.65%. It is clear that data quality and algorithm accuracy are the keys to traffic flow prediction. This study can provide a reference for the development of short-term traffic flow prediction of expressway.