Distributed Approach to Real-Time Control of Complex Signalized Networks

Abstract

Urban traffic characteristics are examined along with the various available current approaches to controlling traffic efficiently at signalized intersections. Fixed-time, central, and distributed real-time control approaches are considered. It is determined that the best method of control varies with network and flow conditions. It is concluded that for smooth traffic and uncomplicated networks that provide good opportunity for progression, a centralized quasi-real-time model such as the split, cycle, and offset optimization technique (SCOOT) is appropriate. However, when faced with significant demand fluctuations or interference from transit or emergency vehicles, a distributed real-time model will be required. The paper recommends use of a fully distributed model to deal with such conditions, at least for the foreseeable future until appropriate hierarchical models can be developed.