Incident prediction as a part of incident management.

Many incident management programs have been implemented over the last years, aiming to reduce congestion due to incidents through the quick removal of the incidents from the road. The best way to minimise delays, however, is to prevent incidents from occurring at all, by actively monitoring the conditions on the network and applying traffic management measures when conditions get too dangerous. The Incident Prediction Module as described in this paper collects real-time data on the traffic and weather conditions, and combines this with roadway geometry variables to calculate the incident probability for a road section. The module also offers the possibility to determine what variables cause the high incident probabilities and provides suggestions as to what traffic management measures can improve conditions.