AOID: adaptive on-line incident detection system

The provisions of any emergency management system with respect to the public safety necessitates the inclusion of the transportation network. The transportation network provides a means for mitigation strategies for any disaster, whether it is natural or human-induced. In this paper, we introduce a set of tools to integrate with a traffic information system to provide automatic traffic incident detection and traffic forecast. Current automated incident detection techniques may not perform well under changing traffic patterns, recurrent congestions, and may require large amounts of training data. We propose a solution to mitigate these shortcomings by utilizing predicted traffic models and performing comparative analysis against observed traffic patterns to automatically detect incidents