A Novel Adaptive Traffic Light Control in Connected Vehicle Environment

Maximizing traffic flow at signalized intersections using adaptive signal control (ASC) strategy has proven highly effective in mitigating the risk of traffic congestion. However, existing ASC strategies typically consider the traffic conditions of a single downstream road at a time, limiting their overall effectiveness. To address this limitation and enhance ASC strategies, this paper proposes an improved adaptive signal control (IASC) which considers traffic conditions of multiple downstream roads during phase selection. Furthermore, with the emergence of vehicular ad hoc network (VANET) technologies, traffic light agents can now communicate with vehicle agents. By leveraging this connectivity, this paper proposes a unified framework called VR-IASC which combines vehicle rerouting strategy with the IASC strategy. First, the proposed IASC considers the intentions of the vehicles in the connected vehicle (CV) environment during the selection of the traffic light phases. Then, the road congestion levels are predicted by considering the vehicle flow affected by the selected traffic light phases. Upon detecting the congested roads, alternative routes based on travel time and remaining road capacity are assigned to the vehicles before they enter the congested roads. Based on simulation results, the proposed IASC outperforms the state-of-the-art ASC strategy by 38.1% and 12.3% in terms of percentage reduction in average travel time in a simple grid network and a real-life Kuala Lumpur (KL) network, respectively. Furthermore, by combining the proposed IASC strategy with vehicle rerouting strategy, the proposed VR-IASC framework achieves substantial reductions in average travel time of 42.26% and 34.74% in the grid network and the KL network, respectively.