A Novel Correntropy-Based Estimation Algorithm for Multi-Sensor Systems Under Hybrid Cyber-Attacks Using Multi-Dimension-Student's-t-Kemel

We consider the state estimation of the non-Gaussian systems observed by multi-sensors subject to hybrid cyber-attacks. The wireless communication network facilitates the information exchange while introducing the risk of cyber-attacks. The estimation performance of the conventional estimators no lo...

Full description

Saved in:
Bibliographic Details
Published in2024 6th International Conference on Electronic Engineering and Informatics (EEI) pp. 58 - 62
Main Authors Wang, Guoqing, Zhu, Zhaolei, Ma, Lei, Yang, Chunyu
Format Conference Proceeding
LanguageEnglish
Published IEEE 28.06.2024
Subjects
Cyber-attacks
Cyberattack
Informatics
Information exchange
Kalman filters
Kernel
Multi-kernel correntropy
Next generation networking
Noise
Robust Kalman filter
State estimation
Student's t kernel function
Target tracking
Wireless communication
Online AccessGet full text

Cover

More Information
Summary:We consider the state estimation of the non-Gaussian systems observed by multi-sensors subject to hybrid cyber-attacks. The wireless communication network facilitates the information exchange while introducing the risk of cyber-attacks. The estimation performance of the conventional estimators no longer holds under the coupling effect of cyber-attacks and non-Gaussian noise. To solve the issue, we first introduce a new correntropy defined using the multi-dimension-student' (MDSTK) inspired by the property of maximum correntropy criterion (MCC) in dealing with heavy-tailed noise. Derived according to the newly MDSTK-based cost function, we obtain a novel correntropy-based estimation for the fusion of multi-sensors with non-Gaussian noise under both denial of service and deception attacks. Simulations of a target tracking example are provided to showcase the effectiveness of the proposed algorithms as compared with related works.
DOI:10.1109/EEI63073.2024.10696634