Short-Term Traffic Flow Forecasting of Road Network Based on Spatial-Temporal Characteristics of Traffic Flow

Abstract

This paper has presented a novel approach designed to realize multi-section short-term traffic flow synchronization forecasting in terms of road network. First, the road network is split into sub networks in accordance with traffic flow spatial-temporal characteristics. Second, chaos analysis method is proposed to forecast short-term traffic flow. Short-term traffic flow characteristics have been figured out by the phase space reconstruction technology and G-P algorithm. By analyzing the traffic flow database, the chaos characteristics of short-term traffic flow time series are correspondingly obtained. Furthermore, Elman neural network in which the input is reconstructing time series has been employed to achieve multi-section forecasting. In addition, an empirical study has been carried out to illustrate this approach. Consequently, this approach has been verified by using traffic flow field data on the road network. The results which support the use of this approach is indicating higher accuracy in short-term traffic flow forecasting.