Abstract
This paper proposes a novel resilient control framework for leader-follower multi-agent systems (MASs) when agents’ sensors are compromised by sinusoid false-data-injection (FDI) attacks. Generally, MASs present various attack surfaces in their communication channels and malicious impact by such persistent attacks can propagate through inter-agent communication, potentially disrupting MAS consensus. We propose an observer-based resilient controller by devising an FDI attack estimator to tackle this issue. The attack estimates are then directly integrated into the agent’s state observer, enabling the MAS to maintain resilience under FDI attacks. We present sufficient conditions for the stability of the proposed resilient control framework via Lyapunov stability with the H∞ control criterion and its corresponding linear matrix inequalities (LMIs). Finally, we provide an illustrative MAS example to demonstrate the efficacy of the proposed attack mitigation strategy.
Funding Information
  • NASA University Leadership Initiative (80NSSC20M0161)

This publication has 17 references indexed in Scilit: