Database-centered architecture for traffic incident detection, management, and analysis

This paper presents various issues related to the development of an integrated software architecture for a traffic incident monitoring, mitigation, and analysis system. The novel concept of using a set of distributed databases, having many different functions and types, is proposed for distributed coordination of sensing, control and analysis algorithms. This coordination paradigm using databases makes the whole architecture robust by providing means to efficiently manage current and past states of the monitored environment and the monitoring system. Use of a semantic event/activity database in the integrated architecture also provides high level abstractions through its query language to model traffic incidents and traffic behaviors. We also present experimental results of the use of a concrete database-centered architecture and algorithms in identifying important traffic flow events (such as tail-gating, exit from a ramp, etc.)