Knowledge-based alarm correlation in traffic monitoring and control

Abstract

The paper shows how a knowledge-based approach can be followed to integrate alarm correlation functionality into traffic monitoring and control systems. To this end, the Alarm Correlation Module (MCA) integrated in the System for Automatic Monitoring of Traffic (SAMOT) is described. According to traffic operator expertise and knowledge, the MCA analyzes, filters and correlates traffic flow anomalies detected by standard video image processing boards. As a consequence of anomaly correlation, the MCA creates adequate image sequences to be shown on an operator's closed-circuit TV and displays adequate messages on variable message panels to keep motorists informed. The MCA knowledge base implements a model of traffic flow concerning the most relevant traffic patterns and taking into account time and space dependence of detected traffic anomalies. The MCA is a successful example of the knowledge-based approach applied to traffic monitoring and control that, after a 6 months trial period, is now functioning on A7 and A10 Italian highways.