Traffic congestion is an important cause of pollution and economic loss. If unchecked, these problems are expected to increase, especially in dense cities. Urban traffic control and coordination can address this challenge and improve efficiency, fuel consumption, and safety. In this talk, the problem of controlling traffic lights under fixed and adaptive routing of vehicles in urban road networks is considered. Based on multi-commodity back-pressure, originally developed for routing and scheduling in communication networks, we develop algorithms to road networks to control traffic lights and adaptively reroute vehicles in a decentralized manner. These algorithm dynamically control traffic lights based on estimated congestion levels on nearby roads. To validate the modeling assumptions and assess the throughput and delay performance, we have relies on a microscopic traffic simulator (VISSIM). Results demonstrate that the proposed multi-commodity and adaptive routing algorithms provide significant improvement over a fixed schedule controller and a single-commodity based back-pressure control, in terms of various performance metrics, including queue-length, trips completed, travel times, and fair traffic distribution. We also discuss impact of traffic estimation errors.
Copyright. All Rights Reserved.