An agent-based simulation model of dynamic real-time traffic signal controller

Traffic management has a key role within intelligent transportation systems (ITS). An efficient traffic control system leads to less fuel consumption, gas emissions, and transportation delay. The main goal of this study is to minimize the driver waiting time at intersections and avoid traffic jams. To achieve this goal, an adaptive traffic signal controller model was proposed and verified in this work. A multi-agent based adaptive controller for a dynamic vehicular ad-hoc network (VANET) in a single four-way intersection was modelled and simulated using Matlab/Simulink/SimEvents. The inputs of the controller are the position and the speed of the vehicles at each approach. In fact, the need for inter-vehicle communication is eliminated in this work through the process of virtual road segmentation and the deployment of road-side units (RSUs). The output of the controlled is the green-time for each approach, where the green time is calculated based on the traffic density and on the queue length. The proposed model was verified by intensive simulations using four main parameters: green phase time; inter-arrival times; service time; and average waiting time (AWT). The performance, AWT, of the proposed adaptive controller was compared to a fixed-time controller. The simulation results showed that the proposed controller outperformed the fixed-time controller for significant variance (i.e. variance > 10) in the means of inter-arrival times for vehicles approaching the traffic signal.