Resilience of Connected and Autonomous Vehicles to Cyber-Physical Disruptions

Abstract

The rise of Connected and Autonomous Vehicles (CAVs) has spurred research into their impact on traffic congestion, which affects travel time, fuel consumption, emissions, and throughput. Additionally, disruptions like lane closures and cyber infrastructure failures create significant obstacles for reliable and efficient travel. While strategies like high-occupancy vehicle lanes aim to reduce congestion, CAVs offer new possibilities to enhance road network performance. We developed and validated an agent-based simulation model to capture interactions between CAVs, non-CAVs, traffic signals, and road infrastructure in both urban and highway environments. Experiments conducted on selected study sites in Oklahoma showed that CAVs improved travel times, even with 30% communication failures and 20% random lane closures. CAVs effectively managed congestion through redundancy mechanisms during communication disruptions; however, CAVs with active communication performed 5% better than those relying solely on redundancy.