Cyber-security challenges of agent technology in intelligent transportation systems

Abstract

Multi-agent approaches have a key role to play in intelligent transportation systems as such systems must service dynamically changing environments in an autonomous fashion. Such Intelligent Transportation Systems (ITS) are exemplified by Google's driverless car. The driverless car can be considered a multi-agent system because it involves multiple interacting intelligent agents analysing and detecting stimuli in its environment and mobilising suitable and safe responses to those stimuli. This position paper discusses the security threats that may result from introducing multiagent technology in such intelligent transportation systems. We argue that, while multi-agent approaches have a major role to play in this domain, their introduction in vehicle-to-vehicle communication or autonomous driving systems raises new security challenges that must be addressed by both cyber-security and multi-agent research.