Probability-Guaranteed Distributed Set-Membership Secure Fusion Estimation Against Nonlinear Hybrid Attacks

Abstract

This paper investigates the distributed secure fusion estimation problem under stochastic nonlinear hybrid attacks. Specifically, this work analyzes a hybrid attack scenario where the attacker employs a randomized approach to launch false data injection (FDI) attacks and Denial-of-Service (DoS) attacks on the measurement information communication channel. Then, an innovative distributed secure fusion estimation model is proposed, addressing three situations: DoS attacks, FDI attacks, and the absence of attacks. Following this, an existence condition is derived for the secure fusion estimator, utilizing probability-guaranteed set-membership filtering technology, to ensure that the fusion estimation error will consistently be bounded within an expected ellipsoid with the specified probability. Subsequently, a convex optimization problem involving constrained recursive matrix inequalities is formulated to compute the secure fusion estimation weight matrices. Finally, the effectiveness of the proposed probability-guaranteed set-membership secure fusion estimation (SSFE) algorithm is demonstrated through a simulation example. Note to Practitioners—The research in this paper is dedicated to addressing the problem of fusion state estimation in practical engineering tasks such as intelligent transportation, industrial manufacturing and military defense. With the increase in application requirements and process accuracy, the majority of projects demand that the true state must be bounded within a certain range, e.g., unmanned vehicle obstacle avoidance and missile precision strikes. To overcome this challenge, probability-guaranteed set-membership filtering is introduced to ensure that the fusion estimation error is bounded with a certain probability. However, due to the expanding scope of engineering applications, the system may be deployed to perform tasks in a non-secure environment, which increases the risk of malicious attacks that may lead to functional failures as well as performance degradation. Therefore, this paper simultaneously considers the scenario where the communication channels are subjected to the stochastic hybrid attacks, which can be effectively handled by utilizing the proposed probability-guaranteed SSFE algorithm. Preliminary simulations demonstrate the feasibility of the algorithm. Since the actual system may also encounter problems such as sensor energy constraints, bandwidth resource limitations, and data processing asynchrony. Therefore, in our future work, we will focus on addressing the various limitations present in the fusion estimation system and improving their resolution.

Topics & Concepts