Data-Driven Adaptive Iterative Learning Control for Consensus Tracking of Multiagent Systems Subject to Hybrid Cyber...

Area of Science:

• Control Theory
• Cybersecurity
• Networked Systems

Background:

• Multiagent systems (MASs) are crucial for distributed tasks but vulnerable to cyberattacks.
• Hybrid cyberattacks, combining denial-of-service (DoS) and deception, pose significant threats to MAS consensus.
• Existing control methods struggle to address data-driven needs under complex attack scenarios.

Purpose of the Study:

• To develop a data-driven adaptive iterative learning control (ILC) for nonlinear MASs under hybrid cyberattacks.
• To address both connectivity-maintained and connectivity-paralyzed hybrid attack scenarios.
• To provide a robust control strategy that compensates for system uncertainties and cyber threats.

Main Methods:

• A data-driven adaptive ILC approach is proposed, utilizing iterative estimation for compensation.
• The control strategy is extended to handle hybrid attacks that paralyze system connectivity.
• Rigorous mathematical analysis is employed to prove convergence and stability.

Main Results:

• The proposed data-driven adaptive ILC effectively achieves consensus in nonlinear MASs under hybrid cyberattacks.
• The approach demonstrates robustness against both DoS and deception attacks.
• Explicit characterization of performance impact under connectivity-paralyzed attacks is achieved.

Conclusions:

• The developed data-driven adaptive ILC offers a viable solution for robust consensus control in nonlinear MASs facing sophisticated cyber threats.
• The method's effectiveness is validated through numerical simulations, showcasing its practical applicability.
• This work contributes to secure and resilient control of networked systems in adversarial environments.