Lessons Learned and Future Directions for Security, Resilience and Artificial Intelligence in Cyber Physical Systems

Cyber-physical systems (CPS) are used in various safety-critical domains such as robotics, industrial manufacturing systems, and power systems. Faults and cyber attacks have been shown to cause safety violations, which can damage these systems and endanger human lives. The past decade has seen the proliferation of research efforts related to security and resilience in cyber physical systems, with an abundance of publications, workshops, and even media attention. More recently, artificial intelligence (AI) and machine learning (ML) have been reinvigorated and become the topic of paramount attention across the research community, mass media and society, highlighted by trademark successes in the fields of gaming, video, audio, and language translation. While it seems natural to apply AI/ML in CPS security and resilience, the authors would like to share some lessons learned and future directions as cautionary notes, which include: (1) the critical importance of physics and the physical world (P); (2) various means and effects that are introduced by cyber (C); (3) interactions and ramifications of P and C in a system (S); (4) system model and control in CPS; (5) enhanced robustness of control and autonomy in CPS by AI/ML; (6) pitfalls and appropriate positioning of AI/ML in CPS security and resilience; and (7) some challenges and opportunities for research and development.