arXiv:2606.19069v1 Announce Type: cross Abstract: This paper compares the performance of model-free controllers on a nonlinear system under cyberattacks, including false data injection and denial-of-service attacks. Four RL reward types are analyzed for accuracy, cost, and resilience. Results show that the Lyapunov reward offers the best resilience with low tracking error. Exponential mode also provides good trade-offs with acceptable resilience under moderate training conditions. Progressive and linear rewards converge faster but are less robust. RL-MPCs show strong steady-state resilience bu
The increasing integration of AI in critical infrastructure creates an urgent need for robust control systems against sophisticated cyber threats. This paper addresses current vulnerabilities in cyber-physical systems.
This research provides crucial methodologies for enhancing the resilience of cyber-physical systems, which are foundational to modern society and national security, against evolving cyberattacks. The findings improve the stability and trustworthiness of AI-controlled systems.
This research suggests that specific reward types in model-free reinforcement learning can significantly improve resilience in cyber-physical systems under attack, advancing the practical application of secure AI in critical infrastructure. It offers a pathway to more robust control systems.
- · Critical infrastructure operators
- · Cybersecurity firms
- · AI/ML developers in industrial control
- · National security establishments
- · Cyberattackers targeting critical infrastructure
- · Legacy control system providers
- · Sectors reliant on unresilient cyber-physical systems
Improved resilience of industrial control systems using advanced AI techniques will reduce the success rate of cyberattacks.
Increased adoption of similar model-free RL control methods will lead to a new standard for securing AI-driven operational technology.
Nations and organizations with highly resilient cyber-physical systems will gain a strategic advantage in economic stability and defense capabilities.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG