Optimizing a MisInformation and MisBehavior (MIB) Attack Targeting Vehicle Platoons

摘要

Autonomous driving features can mitigate traffic fatalities, create more enjoyable commutes, and increase fuel efficiency. For example, collaborative adaptive cruise control (or platooning) uses sensor-based distance measurement and vehicle-to-vehicle communications to automatically control inter-vehicle spacing. This can have tremendous benefits but is also safety critical. Therefore, it is essential to understand and mitigate potential platooning vulnerabilities. In this work, we design an attack that we call the insider MisInformation and misBehavior (MIB) attack. During this attack, a malicious vehicle uses misinformation, erroneous V2V communications, and misbehavior, erratic driving, to cause predictable, dangerous, behavior. Although this attack can be applied broadly, we use it to design three optimal attacks were an attacker causes a collision without being damaged. Finally, we simulate these attacks and discuss trade-offs in there design parameters.