Traffic volume forecasting of urban highway in Shanghai.

Forecasting traffic volume is an important task in guiding drivers' routes controlling urban highways, and providing real-time transportation information. In this study, neural network models are used to forecasting traffic volume. First the Rescaled Range (R/S) analysis is used to identifying traffic volume data trends, fluctuations and randomness. Then Time-delayed recurrent network is used to forecast the traffic volume in the next 15 minutes. The experiments show that the traffic volume forecasting based on recurrent model has a good performance.