MPC-Based Asynchronous Attack Tolerant Control for Uncertain Markov Jump Cyber-Physical Systems.

Area of Science:

• Cyber-Physical Systems
• Control Theory
• Network Security

Background:

• Cyber-physical systems (CPS) are increasingly vulnerable to sophisticated attacks like Denial-of-Service (DoS).
• Uncertain Markov jump systems introduce complexities in control design due to mode uncertainties.
• Existing Model Predictive Control (MPC) methods face limitations with asynchronous system information and constraint handling.

Purpose of the Study:

• To develop an asynchronous attack-tolerant control scheme for uncertain Markov jump cyber-physical systems (MJCPSs) under DoS attacks.
• To address the challenge of unobserved system running modes at the control center.
• To enhance system performance and reduce computational burden in MPC.

Main Methods:

• An asynchronous model predictive controller is proposed, combining dynamic and state feedback controllers.
• A decision variable is introduced to overcome feasible region limitations in traditional MPC.
• Solvable optimal problems are formulated to achieve desired system performances.

Main Results:

• The proposed asynchronous MPC scheme effectively tolerates DoS attacks in uncertain MJCPSs.
• The dynamic controller allows online performance tuning and expands the initial feasible region.
• The method alleviates online computation burden and overcomes traditional MPC constraints.

Conclusions:

• The developed asynchronous MPC provides a robust solution for attack-tolerant control in uncertain MJCPSs.
• The controller enhances system resilience against DoS attacks while managing computational complexity.
• The proposed approach demonstrates practical applicability through a successful application example.