Vulnerability assessment of large-scale urban road network under emerging disinformation attacks

Abstract

Transportation Cyber-Physical Systems (TCPS) are facing an emerging cyber threat from disinformation attacks that can manipulate individual cognition and behaviours, potentially making urban transportation vulnerable. Traditional studies on road network vulnerability have primarily focussed on physical attacks like node or link removal, with little attention to disinformation attacks. To fill this gap, this study introduces a novel disinformation attack mode for road TCPS. It hacks navigation applications by strategically modifying the link cost information, thereby affecting drivers' routing decisions. Crucially, this attack mode leaves the physical topology of TCPS-based road networks unchanged, impacting only the cyber layer's information. Additionally, intriguing link metrics, like the partial derivative of total travel time to link flow, are introduced to identify attack targets. On this basis, we design multi-strategy disinformation attacks to assess road network vulnerability. The proposed research framework, validated by San Francisco's large-scale urban road network, reveals that disinformation attacks on just 0.12% of links could cause an annual city-wide economic loss of $79.26 million in the worst-case scenario. This study offers a unique viewpoint on road network vulnerability, emphasising the vital need for TCPS cybersecurity to ensure urban transportation's reliability and resilience.