Security vulnerabilities in the Connected and Autonomous Vehicle (CAV) software can allow hackers to perform malicious actions ranging from draining batteries and taking control of the steering wheel to disabling the alarm system. CAV software is subject to several cyberattacks, including memory corruption, code injection, remote code execution, and malware attacks. This project aims to develop a Network Functions Virtualization (NFV)-based virtualized security framework to enhance the resilience of Connected and Automated Vehicle (CAV) software. The proposed framework integrates a reinforcement learning (RL)-based code diversification mechanism to optimally execute code variants of CAV software, which is implemented as virtual network functions (VNFs). Our RL agent can choose the optimal policy (i.e., the optimal code variant) that maximizes the resilience, ensures quality of service (e.g., service uptime), and reduces the cost incurred in executing VNFs. By meticulously switching the VNFs, the framework presents an unpredictable attack surface, thereby preventing hackers from exploiting vulnerabilities.
This project was funded by the National Center for Transportation Cybersecurity and Resiliency (TraCR)