Core Concepts
Secure consensus in multi-agent systems under FDI attacks is achieved through passivity-based attack detection and switching controllers.
Abstract
The paper addresses output consensus in linear passive multi-agent systems under FDI attacks. It proposes a passivity-based approach for detecting attacks and a switching controller for defense mode. The event-triggered observer feedback ensures system stability. Simulation examples support theoretical findings.
I. Introduction
Cyber-Physical Systems (CPS) face FDI attacks on actuation signals and sensor measurements.
Event-triggered cryptographic authentication enhances security with limited resources.
II. System Modeling and Problem Description
Agents interact over a network susceptible to malicious FDI attacks.
III. Event Condition and Zeno Behavior
An event condition ensures observer stability without Zeno behavior.
IV. Attack Detection, Mitigation, and Closed-Loop System Stability
Passivity-based attack detection relies on verifying passivity inequality.
Switching controllers mitigate attacks for achieving practical output consensus.
V. Simulation Example
A. Passive Agents with Real Poles
Passivity lost during attack detected by the controller switching to defense mode.
B. Passive Agents with Complex Poles
Oscillations affect attack estimation performance of the system.
C. Passive Agents with Imaginary Poles
Higher event density due to oscillating nature of output affects system performance.
VI. Conclusions
Proposed method achieves secure consensus by detecting and mitigating attacks effectively.
Stats
Passive agents are characterized by positive real transfer functions ensuring passivity lemma holds true.