Toward an Autonomic Architecture for Real-Time Traffic Network Management

This paper presents an autonomic-based architecture for real-time traffic management in congested urban transportation networks. The architecture assumes the availability of a spatially distributed set of local controllers in the network. Each controller is capable of monitoring the traffic within a predefined subnetwork and provides efficient control strategies for its traffic. Following a predefined communication protocol, controllers are assumed to be able to share information on the observed traffic pattern and their control actions. In addition, controllers could be dynamically configured to operate in teams to develop integrated traffic management schemes that best cope with the observed traffic pattern in the network. This paper presents the results of an off-line testing of the proposed architecture. It investigates the most efficient team-formation strategies among controllers to mitigate a non-recurrent traffic congestion situation in a typical highway network. The results show that more efficient traffic management strategies could be obtained through collaboration among individual controllers, which result in considerable travel time savings. BACKGROUND Metropolitan areas in the U.S. and abroad are suffering severe traffic congestion problems that are depicted by considerable travel delays, poor environmental quality, high accident rates, and increasing traveler frustration (Schrank & Lomax 2007). With the limited ability to expand the physical capacity of the existing transportation system to meet the growing demand, there are increasing calls for the development of innovative technologies to improve the system's throughput, efficiency and safety. Intelligent Transportation Systems (ITS) have emerged over the last few decades as one plausible approach. They integrate advances in communications, surveillance, and computational technologies to provide real-time traffic management capabilities for congested urban networks. Thus, considerable investment has been allocated to furnish the transportation infrastructure with hardware and software technologies that support various ITS services. In addition, Traffic Management Centers (TMCs) have been established in many cities to oversee the use of these technologies to alleviate traffic congestion. Nonetheless, traffic network management efforts at most TMCs has been limited to using traffic surveillance technologies to monitor network conditions and to warn of any detected irregularities (e.g., roadway incidents). In most cases, network managers rely on their experience to evaluate the impact on network performance and to decide if a traffic management scheme can be deployed. A scheme is typically restricted to warning travelers of delays associated with an incident, and providing them with primitive route diversion recommendations that most likely do not guarantee the optimal use of the …