Intelligent transportation systems that lack reliable and accurate traffic condition prediction are, at best, reactive. Although rapid response to current conditions is a significant improvement over past transportation systems, ITS in the new millenium must include proactive system control and information delivery. Research in traffic condition forecasting has intensified in response to this need. However, a broad consensus on preferred forecast methods has yet to emerge. There is now a need to test the promising prediction methods in an operational setting. The methods that prevail must be both accurate and easy to implement on a large scale. Two promising methods, seasonal time series models and non-parametric regression, are being evaluated in ongoing research using real time freeway data at Virginia's Smart Travel Laboratory. (A*)
Samenvatting