Fault-tolerant Reconfigurable Control for Mimo Systems Using Online Fuzzy Identification

In this paper, a novel fault-tolerant control scheme is proposed for a class of uncertain multiple-input-multiple-output (MIMO) nonlinear systems based on online fuzzy clustering and identification. The unknown nonlinear function of the system is identified online using an evolving Takagi-Sugeno (T-S) fuzzy model and the fault-tolerant control law is designed based on the identified fuzzy model. The proposed method has the following features: (i) both the structure and the parameters of the T-S fuzzy model can evolve online, which makes it capable of representing more parameter and structure uncertainties of the nonlinear system; (ii) an online fuzzy clustering algorithm is employed to determine the generation of a new cluster center (new rule), which also serves as a warning signal of detecting a fault; (iii) a self-structuring fault-tolerant controller is constructed based on the online identified fuzzy model, together with a compensation controller, to guarantee the closed-loop stability under parameter variations and faults. The proposed fault-tolerant control scheme does not rely on a specially designed fault diagnosis module. Simulation studies on an inverted pendulum example have verified the effectiveness of the proposed control scheme and demonstrated the proposed control scheme has the capability to achieve desired tracking performance when there exist parameter changes or faults and the generation of new rules can alert the occurrence of faults in the system.