Fuzzy-model-based fault detection for nonlinear networked control systems with periodic access constraints and Bernoulli packet dropouts

• Published in: Applied soft computing Vol. 80; pp. 465 - 474
• Main Authors: Baigzadehnoe, Barmak, Rezaie, Behrooz, Rahmani, Zahra
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.07.2019
• Subjects: Access constraints and packet dropouts; Fuzzy-model-based fault detection; Networked control system; Residual evaluation; Networked control system; Access constraints and packet dropouts; Fuzzy-model-based fault detection; Residual evaluation
• ISSN: 1568-4946; 1872-9681
• DOI: 10.1016/j.asoc.2019.04.023

In this paper, the problem of fault detection for nonlinear networked control systems subject to both communication constraints and packet dropouts is investigated. Despite the huge bulk of studies in the area of linear networked control systems, nonlinear networked control systems have been less involved. In the proposed method, Takagi–Sugeno fuzzy model is employed to represent nonlinear networked control systems. To model the communication network, access constraints and packet dropouts are respectively considered as periodic sequences and Bernoulli processes. Based on the optimal periodic residual generator theory, a fuzzy-model-based observer is proposed to deal with the problem of fault detection for nonlinear networked control systems. Both residual evaluation and correspondence threshold are also investigated to generate confident fault alarms. Finally, the effectiveness of the proposed fault detection scheme for nonlinear networked control systems is illustrated and some simulation results are carried out to validate the proposed approach. •A novel FD scheme for nonlinear model of NCS is proposed.•Both access constraints and packet dropouts in the model of NCS are considered.•T–S fuzzy model is utilized to represent nonlinear NCS.•A fuzzy-based observer is also proposed to deal with the problem of nonlinear FD.•Residual evaluation & correspondence threshold are used to generate fault alarms.