The development of advanced traffic-management systems as a major comp
onent of intelligent vehicle/highway systems is targeted to address th
e complex problem of urban traffic congestion. The strategy includes r
eal-time diversion of traffic during congestion-causing events, by mak
ing efficient use of the available capacities in the highway network.
A conceptual framework for executing real-time diversions is presented
in this paper. The strategy envisages on-line control and management
of traffic through control centers and their surveillance equipment. T
he approach relies on real-time traffic data to detect incidents and d
evelop rerouting plants to benefit motorists. For this purpose, a mode
l has been developed to estimate traveler destinations from traffic da
ta. The dynamic nature of the problem is captured through a traffic as
signment/simulation model, which can predict traffic conditions that w
ould prevail in the alternate routes. A knowledge-based expert system
is also proposed to estimate traffic delays, and to evaluate and choos
e alternate routes. This research is being applied to the northern Vir
ginia urban area as a case study.