Skip to main navigation Skip to search Skip to main content

Resilient output regulation in heterogeneous networked systems under Byzantine agents

  • Jiaqi Yan
  • , Chao Deng*
  • , Changyun Wen
  • *Corresponding author for this work
  • Nanyang Technological University
  • Nanjing University of Posts and Telecommunications

Research output: Contribution to journalArticlepeer-review

Abstract

In this paper, we consider the problem of output regulation in heterogeneous networked systems. In order to cooperatively achieve the global objective, each agent is required to share information in its neighborhood through communication channels, which might be exposed to malicious cyber attacks. Past work shows that the performance of cooperative control can be seriously affected by network misbehaviors. Inspired by such security concerns, this paper considers the resilient cooperative output regulation (COR) in adversarial environment. Despite the presence of so-called Byzantine agents, the normally behaving ones still aim to achieve the goal of trajectory tracking or disturbance rejection. Towards this end, resilient observers with low computational cost are first developed based on a continuous-time resilient consensus protocol. When the number of Byzantine nodes is locally upper bounded and the network structure satisfies certain robustness properties, the designed observers enable each benign agent to exponentially track the exogenous signal. After that, decentralized controllers are proposed facilitating the achievement of resilient COR in the heterogeneous networks. A numerical example is finally given to illustrate the effectiveness of our approach.

Original languageEnglish
Article number109872
JournalAutomatica
Volume133
DOIs
StatePublished - Nov 2021
Externally publishedYes

Keywords

  • Byzantine agents
  • Cooperative output regulation
  • Resilient algorithms

Fingerprint

Dive into the research topics of 'Resilient output regulation in heterogeneous networked systems under Byzantine agents'. Together they form a unique fingerprint.

Cite this