Distributed filtering for nonlinear systems under false data injection attack

This paper investigates a distributed secure estimation problem for a nonlinear stochastic system subject to false data injection attack. The target plant is disturbed by unknown input with non-prior information distribution. Considering the limited communication resources and potential malicious attack in wireless sensor network, a novel protector is designed for each sensor to resist the hostile attack, in which a dynamic decision rule is developed to further reduce the data receiving frequency, thus saving computing resources. Sufficient conditions are established to ensure that the estimation error is exponentially bounded in mean square. Furthermore, the critical attack probability when the steady-state estimation error exceeds the preset value is discussed in two cases. Finally, effectiveness of the proposed technique is demonstrated by a numerical example. (C) 2022 Elsevier Ltd. All rights reserved.

Automated summary

This paper investigates a distributed secure estimation problem for a nonlinear stochastic system subject to false data injection attack. A novel protector is designed for each sensor to resist hostile attack, and a dynamic decision rule is developed to reduce data receiving frequency. The paper establishes sufficient conditions to ensure bounded estimation error and discusses the critical attack probability for the steady-state estimation error exceeding a preset value.