IMPROVED FREEWAY INCIDENT DETECTION USING FUZZY SET THEORY

Freeway incidents often occur unexpectedly and cause undesirable traffic congestion, mobility loss, and environmental pollution even where computerized traffic management systems are installed and in operation. Automatic incident detection, being one of the primary functions of computerized freeway traffic management systems, must be able to detect all freeway incidents as soon as possible with minimum false alarms. In a study that evaluated the applications of fuzzy set theory to improve existing incident detection algorithms, the potential system performance was compared with that of conventional systems using real-world volume and occupancy data that were collected earlier. The potential benefits and needed improvements in the existing incident detection algorithms to take advantage of the promising fuzzy set methodology are summarized.