Multi-agent based traffic light management for privileged lane

Abstract

To optimize the travel time of priority vehicles (PV) and fluidize road traffic in Moroccan cities, we propose in this paper a control system to manage traffic at signalized intersections in urban areas, based on a multi-agent technology and fuzzy logic. the traffic system flow is divided into two types of vehicles; priority and non-priority vehicles, and a group of lanes assigned to each type, the ordinary vehicles can exploit only the ordinary lanes, while the PV may use both the priority and the ordinary lanes. This approach aims to favor emergency vehicles and promote collective modes of transport, it acts on the durations of the traffic light phases to control the different flows. We proposed a decentralized system of regulation based on the supervision of each lane which connects two intersections to build a local view of the crossroad, and intelligent cooperation between neighboring intersections to develop an overview of the environment. The regulation and prioritization decisions are made in real-time by fuzzy logic inference, communication, collaboration, and coordination between different agents. The performance of the proposed system is validated and designed in ANYLOGIC simulator, using a virtual road network that represents a section of the Marrakesh network.