Agent-based Mesoscopic Urban Traffic Simulation based on Multi-lane Cell Transmission Model

Abstract

With the increasing demand for the urban traffic simulation, the mesoscopic model takes center stage with its advantage that can simulate a large network with less computational cost and gets the information of individual vehicles. This study proposes a mesoscopic traffic simulation model called Agent-based Mesoscopic Cell Transmission Model (AMCTM). It models both longitudinal and lateral movement of vehicles. The longitudinal movement is based on the Cell Transmission Model (CTM) and implements the traffic signal by adjusting the amount of flow between cells according to the signal. The lateral movement includes mandatory lane-change to allow the vehicle to follow the specified route and the optional lane-change to reduce the travel time. To evaluate the proposed model, we conduct the experiment with a test network including signalized intersection. The result of the experiment shows that the AMCTM performs well several phenomena occur in the urban network. This study is expected to contribute to the performance evaluation of the newly developed traffic-related technologies such as autonomous vehicles and V2X communication-based control systems.