A method of short term traffic flow prediction that based on the time series theory
COMPUTER MODELLING & NEW TECHNOLOGIES 2014 18(12C) 125-129
School of Civil Architecture, East China Jiaotong University, Nanchang 330013, P.R. China
Road flow is an important base data for traffic control and management, especially, short term future traffic flow is a critical parameter for dynamic travel induction and its control. Many methods have been developed for the problem. But previously models have some deficiency, such as bad adaptability, large amount of calculation needing and many history data requirement. The purpose of this study is to develop a model that can estimate traffic flow on road using the theory of time series treatment and prediction. Karhunen-Loeve transform and spectral analysis have good performance in time series evaluation, describes the method that applied the function of Karhunen-Loeve transform to decompose the history detection traffic flow data series, at the same time get the eigenvector coefficients, use the coefficients and current detection flow reconstructed the future some step traffic flow series, so get the goal of short term traffic prediction. The case study suggest that, the proposed method has a good performance on the prediction, furthermore, a fewer history data needing and several step can be predicted.