Human-Centric Cyber Threat Intelligence for Autonomous Vehicle Networks

Abstract

As the rise in interest in the application of integrated autonomous capabilities to Commercial Autonomous Vehicles (CAVs), e.g., robots mobile drones, and especially, Private Autonomous Vehicles, such as autonomous trucks, autonomous trains, and autonomous passenger vehicles, known as autonomous vehicles (AVs), has reached unprecedented levels, we have noted the increasing importance of intelligent security and privacy architectures, ranging from strategies to promote clean-slate techniques to ensure robustness, security, and privacy of AVs. However, many of these security and privacy concerns are far from solved. Furthermore, as ubiquitous AI technologies appear, new challenges must be addressed and not just near term concerns. Although the capability to perceive the environment and to act on this understanding represents a major technological leap, it fails to solve the privacy concerns linked with always-on sensors, or the security threats posed by adversarial challenges to the vehicle's AI at design, test, or operation stages, as well as during handover or back-transition stages.