Connected Autonomous Vehicle Control Optimization at Intersections

Abstract

Connected and Autonomous Vehicle (CAV)-enabled traffic system has demonstrated great potential to mitigate congestion, reduce travel delay, and enhance safety performance. According to the U.S. Based on seamless Vehicle-To-Vehicle (V2V) and Vehicle-To-Infrastructure communication as well as autonomous driving technologies, traffic management and control will be revolutionized. The existing studies indicate that traffic lights will be eliminated and 75 % of vehicles will be autonomous vehicles by 2040. National Highway Traffic Safety Administration (NHTSA) plans to mandate inter-vehicle communication technologies on every single vehicle by 2016. However, one should note that the current research regarding CAV system management and control is still in its early stage. The presented study concentrates on the VISSIM-based simulation platform development to enable an innovative autonomous intersection control mechanism and optimize CAV operations at intersections without signal lights. Simulation-based investigation on traffic system operations provides a cost-effective, risk-free means of exploring optimal management strategies, identifying potential problems, and evaluating various alternatives. In the study, a VISSIM-based simulation platform is developed for simulating individual-CAV-conflict-based traffic control optimization at intersections. A novel external module will be developed via VISSIM Component Object Model (COM) interfaces. A new CAV-based control algorithm entitled a Discrete Forward-Rolling Optimal Control (DFROC) model, is developed and implemented through the VISSIM COM server. This external module can provide sufficient flexibility to satisfy any specific demands from particular researchers and practitioners for CAV control operations. Research efforts will be made to calibrate driving behavior parameters in the simulation model using drivers’ characteristic data to further strengthen the simulation creditability. Furthermore, a method for statistically analyzing simulation outputs and examining simulation reliability is developed. The methodology developed is applicable for quantitatively evaluating the impacts of various CAV control strategies on urban arterials.