Decentralized trip-based traffic signal control

Abstract

This article proposes a traffic signal control strategy that incorporates the remaining travel distance of each vehicle in the network to clear vehicles faster, reduce congestion, and increase network throughput. The objective of the proposed strategy is to prioritize vehicles with shorter remaining travel distances. By swiftly clearing these vehicles, we free up space for other traffic and alleviate pressure on congested areas. This approach introduces a computational burden due to the need to process individual vehicle data in real-time. To address this, a decentralized, pressure-based control strategy is designed that operates at each intersection and relies on local information. The pressure-based controller dynamically adjusts phase durations based on local phase pressures, maximizing network throughput and stabilizing traffic flows. Simulation results using an AIMSUN model of the Barcelona network, with 1500 links and 600 intersections, show that our approach outperforms strategies based only on links’ queue information.