Resilient Consensus of Multiagent Systems Under Malicious Attacks: Appointed-Time Observer-Based Approach.

Area of Science:

• Control Theory
• Networked Systems
• Cybersecurity

Background:

• Multiagent systems are susceptible to malicious attacks compromising their coordinated behavior.
• Achieving resilient consensus under attacks requires robust state estimation and control strategies.
• Existing methods often demand high communication connectivity, limiting practical applications.

Purpose of the Study:

• To develop an appointed-time observer-based framework for resilient consensus control in linear multiagent systems.
• To address the challenge of unknown malicious attacks on system states.
• To reduce communication requirements for achieving resilient consensus.

Main Methods:

• Design of a local appointed-time state observer for each agent to estimate states accurately despite attacks.
• Proposal of a resilient control strategy using virtual systems and ideal state exchange.
• Investigation of leader-follower resilient consensus with a trusted leader agent.

Main Results:

• The proposed observer effectively estimates agent states at an appointed time under malicious attacks.
• The resilient control strategy facilitates consensus among normal agents through ideal state exchange.
• The framework achieves resilient consensus with significantly reduced communication connectivity requirements (directed spanning tree).

Conclusions:

• The appointed-time observer-based framework provides an efficient solution for resilient consensus control in linear multiagent systems under attacks.
• The leader-follower configuration ensures a specified consensus trajectory while maintaining resilience.
• The reduced communication dependency enhances the practicality and scalability of the proposed approach.