Self-Healing Road Networks: A Self-Organized Management Strategy for Traffic Incidents in Urban Road Networks

Abstract: We propose a novel self-organized traffic management strategy for incidents in traffic-light-controlled road networks. During incidents, it regulates the inflow into already congested road segments by restricting or skipping green times. Simultaneously, the remaining green times are used for the yet free turning directions. In this way, drivers can utilize unused road capacities and circumvent congested areas by revising their originally chosen routes. Consequently, the network has "healed itself" as soon as all affected traffic flows have been redistributed along remaining road capacities. Driver's route choice is modeled by means of discrete choice theory and regards the signaling at the next observable intersection. To this end, we propose a microscopic route choice model for event-oriented decisions. Subsequently, we examine four distinct incident scenarios in a grid and real-size road network and compare fixed-time and traffic-dependent traffic light controls with and without the self-healing strategy employed. As result, the self-healing strategy accounts for significantly lower vehicle accumulations during and after the incident as it effectively antagonizes gridlock effects in the disturbed network.