Host: The Japanese Society for Artificial Intelligence
Name : The 36th Annual Conference of the Japanese Society for Artificial Intelligence
Number : 36
Location : [in Japanese]
Date : June 14, 2022 - June 17, 2022
Traffic congestion, which causes economic loss and environmental pollution, is a serious problem. In order to reduce traffic congestion, numerous researches have been conducted on traffic signal control with appropriate parameters. This study proposes a dynamic offset control method utilizing a multi-agent model for wide-area control applicable to general road networks. In this method, each intersection on the road network acts as an agent, negotiates among the other agents, connects with them, and builds a tree structure to construct an area for dynamic offset control. Inside the tree structure, agents perform green wave control based on the traffic at the edges. We evaluated the proposed method using a grid-like road network and a real road network constructed on a simulation and found that it can dynamically and flexibly form green waves in accordance with traffic conditions, thus leading to an excellent traffic flow facilitation effect.