The study of short-term traffic flow forecasting based on theory of chaos

Traffic flow forecasting has attracted much interest in current literature because of its importance in both the theoretical and empirical aspects of ITS deployment. Many models and methods have been presented in the past. But most of them regard the transportation system as the linear system and using the linear theory to predict the traffic flow. In fact, transportation system is a nonlinear system and traffic flow data exhibits chaotic properties. In this paper, we try to use the chaos theory to forecast the traffic flow in a short-term. Usually there is noise in the collected data which decrease the forecasting precision. So we denoise the data using wavelet transform before forecasting in this paper. The experiment is performed for inductance loop data collected in five minutes interval from the viaduct of Yan'an road in Shanghai in China. And at last our study concludes that techniques based on phase space reconstruction can be used to predict the traffic flow in a short-term. Furthermore, the prediction result is accurate and reliable.