Critic-only adaptive dynamic programming algorithms' applications to the secure control of cyber-physical systems

Area of Science:

• Control Engineering
• Cybersecurity
• Artificial Intelligence

Background:

• Industrial cyber-physical systems (ICPS) are vulnerable to malicious attacks and external perturbations, making security a critical research area.
• Existing security measures often struggle to address complex threats and system dynamics effectively.

Purpose of the Study:

• To propose a novel intelligent secure control scheme for ICPS that integrates optimal control, zero-sum game theory, reinforcement learning, and neural networks.
• To address the challenges posed by actuator attacks and unmatched perturbations in compromised ICPS.

Main Methods:

• The secure control problem is reformulated as a zero-sum game for a nominal auxiliary system.
• Adaptive dynamic programming methods, including policy-iteration and value-iteration, are employed to solve the Hamilton-Jacobi-Isaacs equations.
• Lyapunov stability theory is used to prove the stabilization of compromised systems.

Main Results:

• The proposed scheme effectively mitigates the impact of actuator attacks and unmatched perturbations.
• The intelligent secure control stabilizes compromised cyber-physical systems by optimizing performance parameters.
• Validation on the Quanser helicopter demonstrates the practical effectiveness of the approach.

Conclusions:

• The integrated intelligent secure control scheme offers a robust solution for enhancing the security and stability of industrial cyber-physical systems.
• The methodology provides a framework for addressing complex security challenges in networked control systems.
• The successful application to a physical system highlights the real-world applicability of the proposed techniques.