Robust fault tolerant tracking controller design for unknown inputs T-S models with unmeasurable premise variables

► We study the problem of fault-tolerant control for uncertain TS continuous fuzzy model subject to faults. ► We use a PI observer, subject to unmeasurable decision variables, to estimate jointly states and constant faults. ► We propose LMI design conditions for T-S models tracking control based on descriptor approach. ► The influence of the model uncertainties is minimized with the L2 optimization criterion. ► We apply the derived results to a real process of a wastewater treatment plant.