This study investigates a simple design method of the robust state/fault estimation and fault-tolerant control (FTC) of discrete-time Takagi-Sugeno (T-S) fuzzy systems. To avoid the corruption of the fault signal on state estimation, a novel smoothing signal model of fault signal is embedded in the T-S fuzzy model for the robust H state/fault estimation of the discrete-time nonlinear system with external disturbance by the traditional fuzzy observer. When the component and sensor faults are generated from different fault sources, two smoothing signal models for component and sensor faults are both embedded in the T-S fuzzy system for robust state/fault estimation. Since the nonsingular smoothing signal model and T-S fuzzy model are augmented together for signal reconstruction, the traditional fuzzy Luenberger-type observer can be employed to robustly estimate state/fault signal simultaneously from the H estimation perspective. By utilizing the estimated state and fault signal, a traditional H observer-based controller is also introduced for the FTC with powerful disturbance attenuation capability of the effect caused by the smoothing model error and external disturbance. Moreover, the robust H observer-based FTC design is transformed into a linear matrix inequality (LMI) -constrained optimization problem by the proposed two-step design procedure. With the help of LMI TOOLBOX in MATLAB, we can easily design the fuzzy Luenberger-type observer for efficient robust H state/fault estimation and solve the H observer-based FTC design problem of discrete nonlinear systems. Two simulation examples are given to validate the performance of state/fault estimation and FTC of the proposed methods.

Download full-text PDF

Source
http://dx.doi.org/10.1109/TCYB.2020.3042984DOI Listing

Publication Analysis

Top Keywords

state/fault estimation
24
robust state/fault
20
t-s fuzzy
20
smoothing signal
16
signal model
12
fault signal
12
signal
9
estimation
8
estimation fault-tolerant
8
fault-tolerant control
8

Similar Publications

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!