Short-term traffic flow prediction based on embedding phase-space and blind signal separation

Accurate traffic flow forecasting is one of the important issues for the research of intelligent transportation system (ITS).The capability to forecast traffic flow has been identified as a critical need for dynamic traffic control system. The embedding phase-space theory treats the dynamic evolution of traffic flow as a chaos time series, and this provides the possibility to forecast short-term traffic flow accurately. The theory of blind signal processing is widely used in the area of data mining. Practical historic traffic flow can be regarded as a blind signal mixed of real traffic flow and noise introduced by measurement tools. Blind signal separation is a good method to reduce noise and abstract principal components of historic traffic flow series. This paper proposes an approach based on embedding phase-space and blind signal separation, which enables us to realize the de-nosing and forecasting of the traffic flow synchronously, with another advantage of self-adaptive characteristic.