Abstract
In this paper, observer design for nonlinear systems represented by Takagi Sugeno models (T-S) is investigated. The first main contribution concerns the finite time convergence of the estimations, ensured by an impulsive observer with state updates. The second contribution, lies with taking into account unmeasurable parameters, using the Differential Mean Value Theorem (DMVT) to express the disturbed error dynamics into a Linear Parameter Varying system. The stability conditions are formulated in terms of Linear Matrix Inequalities (LMI). To prove the efficiency of the proposed procedure, applications are performed on a chaotic system. The obtained results are pretty satisfying.