Modelling and numerical simulation of sequencing strategies for connected and autonomous vehicles at signal-free intersections

Abstract

Abstract The signal-free intersections employ connected and automated vehicle technology to manage vehicles passing through the intersection. Due to conflicting traffic flows in opposition directions, a proper sequencing of Connected and Autonomous Vehicles (CAVs) at signal-free intersections becomes critical to impacting intersection traffic performance. Based on an examination of CAV queueing rules under the most common sequencing strategies of First-Come-First-Serve (FCFS) and Longest-Queue-First (LQF), commonly employed as benchmarks for evaluating diverse innovative approaches to signal-free intersections, we propose a Dynamic-Queue-Service (DQS) strategy that is tailored to accommodate high traffic demand. To explicitly elucidate the impact of diverse traffic demand in conflicting directions on the queue uncertainty and stochasticity of CAVs, as well as to investigate how various sequencing strategies influence the equity of CAV traffic at signal-free intersections with regard to CAV queueing dynamics under different strategies, we have developed a double-input traffic queueing model and derived a range of metrics, including the queue length, delay, conditional queue length, and variance of queue length. In addition, for the three strategies, we performed a series of numerical simulations to investigate the queueing process of CAVs at signal-free intersections. Numerical results show that under different levels of traffic demand in the conflicting directions, the FCFS, LQF, and DQS strategies output diverse traffic queueing performances, and the DQS strategy is confirmed to be well-suited for the situation of high traffic demand in both conflicting directions.