Mean-square bounded consensus of nonlinear multi-agent systems under deception attack.

Area of Science:

• Control Systems Engineering
• Networked Systems
• Cyber-Physical Security

Background:

• Multi-agent systems (MAS) are susceptible to deception attacks and various noise types.
• Ensuring reliable consensus under adversarial conditions is crucial for MAS performance.
• Existing methods may not adequately address combined random attacks and nonlinear dynamics.

Purpose of the Study:

• To achieve mean-square bounded consensus for nonlinear MAS under deception attacks and noise.
• To develop an effective state estimator robust to tampered measurements.
• To design a detector for identifying actuator input attacks.

Main Methods:

• A recursive algorithm-based estimator for accurate state estimation.
• A centralized controller integrated with cloud computing.
• An agent-side detector for actuator input attack identification.
• Linear Matrix Inequality (LMI) for deriving consensus conditions.

Main Results:

• Sufficient conditions for achieving mean-square bounded consensus were established.
• An upper boundary for consensus error was derived.
• The proposed methods demonstrated effectiveness in simulation examples.

Conclusions:

• The developed framework ensures robust consensus in nonlinear MAS despite deception attacks and noise.
• The integrated estimator, controller, and detector system enhances system security and reliability.
• The LMI-based approach provides a rigorous mathematical foundation for consensus analysis.