Distributed event-triggered state estimation and fault detection of nonlinear stochastic systems

• Published in: Journal of the Franklin Institute Vol. 356; no. 17; pp. 10335 - 10354
• Main Authors: Zhao, Yue, Shen, Yi
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.11.2019
• ISSN: 0016-0032; 1879-2693
• DOI: 10.1016/j.jfranklin.2018.04.027

•Formulate the stochastic process, nonlinearities and uncertainties for the plant, by the utility of the interval type-2 fuzzy models.•Explore an applicable event-triggering condition under WSNs that facilitates the stability analysis and the solution to solve the parameter coupling problem.•Establish the stability criterion of the resulting distributed fuzzy system and the algorithm solving the filter parametric matrices with less computational complexity.•Design effectual fault detection filtering scheme and evaluation approach for the resulting stochastic system. This paper is devoted to investigate the designs of the event-based distributed state estimation and fault detection of the nonlinear stochastic systems over wireless sensor networks (WSNs). The nonlinear stochastic systems as well as the filters corresponding to the multiple sensors are represented by interval type-2 Takagi–Sugeno (T–S) fuzzy models. (1) A new type of fuzzy distributed filters based on event-triggered mechanism is established corresponding to the nodes of the WSN. (2) The overall stability and performance, that is mean-square asymptotic stability in H∞ sense, of the event-driven fault detection system is analyzed based on Lyapunov stability theory. (3) New techniques are developed to cope with the problem of parametric matrix decoupling for solving the distributed filter gains. (4) Finally, the desired event-based distributed filter matrices are designed subject to the numbers of the fuzzy rules and a series of matrix inequalities. A simulation case is detailed to show the effectiveness of the presented event-based distributed fault detection filtering scheme.