Observer Design for a Class of Takagi-Sugeno Descriptor Systems with Lipschitz Constraints

Applied Mathematical Sciences, Vol. 10, 2016, no. 43, 2105-2120 HIKARI Ltd, www.m-hikari.com http://dx.doi.org/10.12988/ams.2016.64142 Observer Design for a Class of Takagi-Sugeno Descriptor Systems with Lipschitz Constraints Abdellah Louzimi 1,2, Abdellatif El Assoudi 1,2, Jalal Soulami 1,2 and El Hassane El Yaagoubi 1,2 1 Laboratory High Energy Physics and Condensed Matter, Faculty of Science Hassan II University of Casablanca, B.P 5366, Maarif, Casablanca, Morocco 2 ECPI, Department of Electrical Engineering, ENSEM Hassan II University of Casablanca, B.P 8118, Oasis, Casablanca Morocco Corresponding author Copyright c 2016 Abdellah Louzimi et al. This article is distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Abstract This paper deals with the problem of the state estimation for both continuous and discrete time nonlinear descriptor systems described by Takagi-Sugeno (TS) structure with unmeasurable premise variables which satisfying Lipschitz conditions. Indeed, the T-S fuzzy observers are synthesized in explicit forms to estimate exactly the unknown descriptor s states. The idea of the proposed result is based on the separation between dynamic and static relations in the T-S descriptor model. Firstly, the method used to decompose the dynamic equations of the algebraic equation is developed. Secondly, the fuzzy observer designs satisfying Lipschitz conditions and permitting to estimate the unknown states are established for both continuous and discrete time cases. The convergence of the state estimation error is studied by using the Lyapunov theory and the stability condition is given in term of only one Linear Matrix Inequalitie (LMI). Finally, an application to a descriptor model of a heat exchanger pilot process is given in order to illustrate the validity and applicability of the proposed method. Keywords: Takagi-Sugeno systems, descriptor systems, fuzzy observers, Lipschitz condition, unmeasurable premise variables, linear matrix inequality (LMI)

2106 Abdellah Louzimi et al. 1 Introduction The problem of design of fuzzy observer for T-S systems described by ordinary differential or difference equations has been widely studied and is still an active area of research. However, nonlinear descriptor models using T-S fuzzy approach which is largely used in the modeling of nonlinear processes can be used to describe the behavior of many chemical and physical processes [1], [2], [3]. This formulation includes both dynamic and static relations. The numerical simulation of implicit models usually combines a resolution routine of an ordinary (differential or difference) equation together with an optimization algorithm. The purpose of this paper is to design a fuzzy observer for both continuous and discrete time nonlinear descriptor systems described by Takagi-Sugeno (TS) structure with unmeasurable premise variables which satisfying Lipschitz conditions. Recall that, the principle of the T-S approach is to apprehend the global behavior of a process by a set of locals model [4], [5]. The advantage of such approach relies on the fact that once the T-S fuzzy models are obtained, some analysis and design tools developed in the linear case can be used, which facilitates observer or/and controller synthesis for complex nonlinear systems see for example [6], [7] and the references therein. However, many researches on T-S fuzzy state observation and its application for dynamic explicit models in the continuous-time and the discrete-time exist in the literature. We may cite [8], [9], [10], [11], [12], [13]. For nonlinear descriptor systems described by T-S models, several developments exist for continuous-time case [14], [15], [16], [17] and for discrete-time one [18], [19]. Moreover, many other works dealing with observer design in explicit form called Proportional Integral Observer were also proposed for implicit T-S models. These results are based on the singular value decomposition approach and generalized inverse matrix and consider the output matrix without nonlinear terms see for example [20], [21], [22], [23] and many references herein. Notice that, generally an interesting way to solve the various fuzzy observer problems raised previously is to write the convergence conditions on the LMI form [24]. In this paper, the main contribution consists to develop a new design methodology based on the separation between the dynamic and static relations in the T-S descriptor model. More precisely, observer designs for both continuous and discrete time T-S descriptor models with unmeasurable premise variables which satisfying the Lipschitz conditions are proposed. The asymptotic stability of the state estimation error is studied by using the Lyapunov theory and the stability condition is given in term of only one LMI. Besides, the proposed fuzzy observers for both classes of continuous and discrete time T-S descriptor systems are synthesized by only an explicit structure. The paper is organized as follows. The class of fuzzy T-S structures with

Observer design for a class of T-S descriptor systems 2107 unmeasurable premise variables of nonlinear descriptor systems is introduced in section 2. Section 3 presents the main results about fuzzy observer design which satisfying Lipschitz conditions. Both continuous-time and discrete-time cases are investigated. Section 4 treats an illustrative application to show the efficiency of the proposed approach in the observer design of a continuous-time T-S descriptor system. Finally, the conclusion is presented in Section 5. In this paper, in order to simplify the presentation of the main results in the next section, continuous-time fuzzy descriptor systems and discrete-time fuzzy descriptor systems are abbreviated as CFDS and DFDS respectively. Other notations used are fair standard. For example, X > 0 means the matrix X is symmetric and positive definite. X T denotes the transpose of X. The symbol I (or 0) represents the identity matrix with appropriate dimension (or zero matrix). 2 Takagi-Sugeno fuzzy descriptor systems In this paper, the following class of T-S fuzzy descriptor models is considered: Ex (t) = γ i (x(t))(a i x(t) + B i u(t)) (1) y(t) = Cx(t) + Du(t) where for the CFDS case x (t) means ẋ(t) and for DFDS case x (t) represents x(t+1). In (1), x(t) = [X1 T (t) X2 T (t)] T R n is the state vector with X 1 (t) R r is the vector of dynamical (differential or difference) variables, X 2 (t) R n r is the vector of algebraic variables, u(t) R m is the control input, y(t) R p is the measured output. E R n n with rank(e) = r, A i R n n, B i R n m, C R p n, D R p m are real known constant matrices with: ( ) ( ) ( ) I 0 A11i A E = ; A 0 0 i = 12i B1i ; B i = ; C = ( C 1 0 ) (2) A 21i A 22i where constant matrices A 22i are supposed invertible (rank(a 22i ) = n r). q is the number of sub-models. The γ i (x(t)) (i = 1,..., q) are the weighting functions that ensure the transition between the contribution of each sub model: { Ex (t) = A i x(t) + B i u(t) (3) y(t) = Cx(t) + Du(t) They verify the so-called convex sum properties: γ i (x(t)) = 1 0 γ i (x(t)) 1 i = 1,..., q B 2i (4)

2108 Abdellah Louzimi et al. Then, before giving the main results, let us make the following assumption [3]: Assumption 2.1 : Suppose that: (E, A i ) is regular, i.e. det(se A i ) 0 s C All sub models (3) are impulse observable and detectable. In order to investigate the observer design for T-S descriptor system (1), the approach is based on the separation between difference and algebraic equations in each sub model (3) and the global fuzzy model is obtained by aggregation of the resulting sub models. So, using (2), system (3) can be rewritten as follows: X 1(t) = A 11i X 1 (t) + A 12i X 2 (t) + B 1i u(t) 0 = A 21i X 1 (t) + A 22i X 2 (t) + B 2i u(t) y(t) = C 1 X 1 (t) + Du(t) The form (5) for system (3) is also known as the second equivalent form [3]. From (5) and using the fact that A 1 22i exist, the algebraic equations can be solved directly for algebraic variables, to obtain: X 2 (t) = A 1 22iA 21i X 1 (t) A 1 22iB 2i u(t) (6) Substitution of the resulting expression for X 2 (equation (6)) in equation (5) yields the following model: X 1(t) = M i X 1 (t) + N i u(t) X 2 (t) = Q i X 1 (t) + R i u(t) y(t) = C 1 X 1 (t) + Du(t) (5) (7) where M i = A 11i A 12i A 1 N i = B 1i A 12i A 1 Q i = A 1 22iA 21i R i = A 1 22iB 2i 22iA 21i 22iB 2i (8) The weighting functions γ i (x(t)), i = 1,..., q can be rewritten as γ i (x(t)) = γ i (X 1 (t), X 2 (t) = Q i X 1 (t) + R i u(t)) = γ i (ξ(t)) (9) with ξ(t) = [X1 T (t) u T (t)] T. In descriptor form, sub system (7) takes the following equivalent form of sub model (3): { Ex (t) = M i x(t) + Ñiu(t) (10) y(t) = Cx(t) + Du(t)

Observer design for a class of T-S descriptor systems 2109 where ( Mi 0 M i = Q i I ) ; Ñ i = ( Ni R i ) (11) Then, fuzzy descriptor system (1) can be rewritten in the following equivalent form: Ex (t) = γ i (x(t))( M i X 1 (t) + Ñiu(t)) y(t) = C 1 X 1 (t) + Du(t) (12) So, using (4), (9) and (11), system (12) can be rewritten in the following equivalent system: X 1(t) = γ i (ξ(t))(m i X 1 (t) + N i u(t)) X 2 (t) = γ i (ξ(t))(q i X 1 (t) + R i u(t)) y(t) = C 1 X 1 (t) + Du(t) (13) In order to take an observer for system (1), we introduce the following matrices: Z 0 = 1 Z i q (14) Z i = Z i Z 0 where Z 0 = M 0, N 0, Q 0, R 0 and Z i = M i, Ni, Qi, Ri with Z i = M i, N i, Q i, R i. By using (14), (13) takes the following equivalent form: X 1(t) = M 0 X 1 (t) + N 0 u(t) + γ i (ξ(t))( M i X 1 (t) + N i u(t)) X 2 (t) = Q 0 X 1 (t) + R 0 u(t) + γ i (ξ(t))( Q i X 1 (t) + R i u(t)) y(t) = C 1 X 1 (t) + Du(t) (15) 3 Main results In this section, the aim is to suggest a new method for the observer design of T-S descriptor model (1) with unmeasurable premise variables. Based on the transformation of the T-S descriptor system (1) into the equivalent form (15),

2110 Abdellah Louzimi et al. the proposed observer is given by the following equations: ˆX 1(t) = M 0 ˆX1 (t) + N 0 u(t) L(ŷ(t) y(t)) + γ i (ˆξ(t))( M i ˆX1 (t) + N i u(t)) ˆX 2 (t) = Q 0 ˆX1 (t) + R 0 u(t) + γ i (ˆξ(t))( Q i ˆX1 (t) + R i u(t)) ŷ(t) = C 1 ˆX1 (t) + Du(t) (16) where ( ˆX 1 (t), ˆX2 (t)), ŷ(t) and ˆξ(t) denote the estimated state vector, the output vector and the decision variable vector respectively. The observer gain L is the unknown parameter to be determined. In order to establish the conditions for the asymptotic convergence of the observer (16), we define: ( ) ( ) ε1 (t) ˆX1 (t) X ε(t) = = 1 (t) (17) ε 2 (t) ˆX 2 (t) X 2 (t) From systems (15) and (16), the dynamics of the estimation error is given by: where ε 1(t) = Γ 0 ε 1 (t) + δ(t) ε 2 (t) = Q 0 ε 1 (t) + γ i (ˆξ(t)) Q i ε 1 (t) + (γ i (ˆξ(t)) γ i (ξ(t)))( Q i X 1 (t) + R i u(t)) (18) Γ 0 = M 0 LC 1 (19) and with δ(t) = ( M i δ 1 (t) + δ 2 (t)u(t)) (20) { δ1 (t) = γ i (ˆξ(t)) ˆX 1 (t) γ i (ξ(t))x 1 (t) δ 2 (t) = N i (γ i (ˆξ(t)) γ i (ξ(t))) (21) Assumption 3.1 : Assume that the following conditions hold: δ 1 (t) < λ i ε 1 (t) δ 2 (t) < µ i ε 1 (t) u(t) < ρ (22) where λ i, µ i are positives scalars Lipschitz constants and ρ > 0.

Observer design for a class of T-S descriptor systems 2111 Using the Assumption 3.1, the term δ(t) can be bounded as follows: where δ(t) < θ ε 1 (t) (23) θ = (σ( M i )λ i + µ i ρ) (24) where σ( M i ) denote the maximum singular value of the matrix M i. Note that to prove the convergence of the estimation error ε(t) toward zero, it suffice to prove from (18), that ε 1 (t) converges toward zero. Then, the following results can be stated. Theorem 3.2 : Under above Assumptions 2.1 and 3.1, the CFDS (18) is globally asymptotically stable if there exists matrices P > 0, Q > 0 and W verifying the following LMI: ( M T 0 P + P M 0 C1 T W T W C 1 + θ2q 2 P ) P Q The gain L of the observer (16) is computed by: < 0 (25) L = P 1 W (26) Theorem 3.3 : Under above Assumptions 2.1 and 3.1, the DFDS (18) is globally asymptotically stable if there exists matrices P > 0, Q > 0 and W verifying the following LMI: where Σ W C 1 P P M 0 W C 1 0 Q < 0 (27) Σ = M T 0 P M 0 P M T 0 W C 1 C T 1 W T M 0 + θ 2 P + θ 2 2Q (28) The gain L of the observer (16) is computed by: L = P 1 W (29)

2112 Abdellah Louzimi et al. Proof of Theorem 3.2 : Consider the following candidate of quadratic function: V (t) = ε T 1 (t)p ε 1 (t) P > 0 (30) The time derivative of the Lyapunov function (30) along the trajectories of the system (18) is obtained as: V (t) = ε T 1 (t)(γ T 0 P + P Γ 0 )ε 1 (t) + δ T (t)p ε 1 (t) + ε T 1 (t)p δ(t) (31) Lemma 3.4 : For any matrices X and Y with appropriate dimensions, the following property holds for any invertible matrix J: X T Y + Y T X X T J 1 X + Y T JY (32) For Q > 0, by applying Lemma 3.4 and Assumption 3.1, (31) becomes: V (t) < ε T 1 (t)(γ T 0 P + P Γ 0 + P Q 1 P )ε 1 (t) + δ T (t)qδ(t) (33) Taking into account (23), (33) becomes: V (t) < ε T 1 (t)(γ T 0 P + P Γ 0 + P Q 1 P + θ 2 Q)ε 1 (t) (34) Thus, the estimation error convergence is ensured if the following condition is guaranteed: Γ T 0 P + P Γ 0 + P Q 1 P + θ 2 Q < 0 (35) Then from (19), we can establish the LMI condition (25) of Theorem 3.2 by using the Schur complement [24] and the following change of variables: W = P L (36) Thus, from the Lypunov stability theory, if the LMI condition (25) is satisfied, the system (18) is exponentially asymptotically stable. This completes the proof of Theorem 3.2. Proof of Theorem 3.3 : Consider the following candidate of quadratic function: V (t) = ε T 1 (t)p ε 1 (t) P > 0 (37) The variation of V (t) along the trajectory of (18) is given by: V (t) = V (t + 1) V (t) (38)

Observer design for a class of T-S descriptor systems 2113 From (18), we have: V (t) = ε T 1 (t)(γ T 0 P Γ 0 P )ε 1 (t) + δ T (t)p Γ 0 ε 1 (t) +ε T 1 (t)γ T 0 P δ(t) + δ T (t)p δ(t) For Q > 0, by applying Lemma 3.4 and Assumption 3.1, (39) becomes: V (t) = ε T 1 (t)(γ T 0 P Γ 0 P + Γ T 0 P Q 1 P Γ 0 )ε 1 (t) +δ T (t)qδ(t) + δ(t) T P δ(t) Taking into account (23), (40) becomes: (39) (40) V (t) = ε T 1 (t)(γ T 0 P Γ 0 P + Γ T 0 P Q 1 P Γ 0 + θ 2 P + θ 2 Q)ε 1 (t) (41) Thus, the estimation error convergence is ensured if the following condition is guaranteed: Γ T 0 P Γ 0 P + Γ T 0 P Q 1 P Γ 0 + θ 2 P + θ 2 Q < 0 (42) Then from (19), we can establish the LMI conditions (27) of Theorem 3.3 by using the Schur complement [24] and the following change of variable: W = P L (43) Thus, from the Lypunov stability theory, if the LMI condition (27) is satisfied, the system (18) is exponentially asymptotically stable. This completes the proof of Theorem 3.3. 4 Application to a heat exchanger system In this section, the proposed fuzzy observer design (16) is applied to a heat exchanger pilot process in order to estimate on-line these unknown states. The following CFDS with unmeasurable premise variables that we consider here is given in [16]. It takes the form: 4 Eẋ(t) = γ i (x(t))(a i x(t) + Bu(t)) (44) y(t) = Cx(t) where x(t) = (x 1 (t),, x 8 (t)) T is the state vector, u(t) = (u 1 (t), u 2 (t)) T is the control vector, y(t) = (x 1 (t), x 4 (t)) T is the vector of output measurements. 0.29 4.14 10 4 0 0.29 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0.10 0 0 0.10 1.47 10 A 1 = 3 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 39.48 8.80 0 0 0 1 0 0 0 0 0 39.48 8.80 0 1

2114 Abdellah Louzimi et al. A 2 = A 3 = A 4 = 0.29 4.14 10 4 0 0.29 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0.10 0 0 1.48 1.47 10 3 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 39.48 8.80 0 0 0 1 0 0 0 0 0 39.48 8.80 0 1 8.57 4.14 10 4 0 0.29 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0.10 0 0 0.10 1.47 10 3 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 39.48 8.80 0 0 0 1 0 0 0 0 0 39.48 8.80 0 1 8.57 4.14 10 4 0 0.29 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0.10 0 0 1.48 1.47 10 3 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 39.48 8.80 0 0 0 1 0 0 0 0 0 39.48 8.80 0 1 E = 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 C = The weighting functions are given by:, B = ( 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 36.72 0 0 36.72 γ 1 (x(t)) = (α 1 α 2 x 2 (t))(α 3 α 4 x 5 (t))/α 5 γ 2 (x(t)) = (α 1 α 2 x 2 (t))(α 6 + α 4 x 5 (t))/α 5 (45) γ 3 (x(t)) = (α 7 + α 2 x 2 (t))(α 3 α 4 x 5 (t))/α 5 γ 4 (x(t)) = (α 7 + α 2 x 2 (t))(α 6 + α 4 x 5 (t))/α 5 )

Observer design for a class of T-S descriptor systems 2115 with α 1 = 8.29, α 2 = 552.59, α 3 = 1.38, α 4 = 92.10, α 5 = 11.45, α 6 = 9.21 10 9, α 7 = 5.53 10 8 Note that, the application of the proposed observer (16) for heat exchanger pilot process requires that the above model (44) takes the form (15). To do so, considering the following: X 1 (t) = [x 1 (t) x 2 (t) x 3 (t) x 4 (t) x 5 (t) x 6 (t)] T, X 2 (t) = [x 7 (t) x 8 (t)] T ( ). I 0 E = with rank(e) = 6. For i = 1, 2, 3, 4: 0 0 ( ) ( ) A11i A A i = 12i Ai (1 : 6, 1 : 6) A = i (1 : 6, 7 : 8) A 21i A 22i A i (7 : 8, 1 : 6) A i (7 : 8, 7 : 8) ( ) ( ) B1 B(1 : 6, 1 : 2) B i = B = = ; B(7 : 8, 1 : 2) B 2 C i = C = ( ) ( ) C 1 C 2 = C(1 : 2, 1 : 6) C(1 : 2, 7 : 8). This shows that the model (44) is a particular case of the system (1). Hence, system (44) can be rewritten in the following equivalent form: Ẋ 1 (t) = M 0 X 1 (t) + N 0 u + γ i (ξ(t))( M i X 1 (t) + N i u(t)) X 2 (t) = Q 0 X 1 (t) + R 0 u + γ i (ξ(t))( Q i X 1 (t) + R i u(t)) y(t) = C 1 X 1 (t) (46) where M 0, N 0, Q 0, R 0, Mi, Ni, Qi and R i can be calculated using equation (14) and γ i (ξ(t)) = γ i (x(t)) (see (9)). By Theorem 3.2, the following observer gain L is obtained: 3.7991 0.0000 0.0238 0.0000 L = 10 3 0.0000 0.0000 0.0000 3.7990 0.0000 0.0008 0.0000 0.0000 The numerical values of control vector are: u 1k = 0.012 and u 2k = 0.012. Simulation results with initial conditions: x 0 = [65 0.002 0.002 20 0.001 0.001 0.3440 0.3923] T ˆx 0 = [65 0.004 0.001 20 0.002 0.002 0.2739 0.3440] T are given in Figures 1 and 2. These simulation results show the performances of the proposed observer (16) with the gain L where the dashed lines denote the state variables estimated by the fuzzy observer. They show that the observer gives a good estimation of unknown states of the heat exchanger pilot process.

2116 Abdellah Louzimi et al. 5 Conclusion In this paper, a new fuzzy observer design approach based on explicit structure is proposed in order to estimate the state of both continuous and discrete time nonlinear descriptor system described by T-S fuzzy models with unmeasurable premise variables and satisfying Lipschitz conditions. The main idea is based on the separation between dynamic and static relations in the T-S descriptor model. The convergence of the state estimation error is studied using the Lyapunov theory and the existence of condition ensuring this convergence is expressed in term of only one LMI. To illustrate the proposed methodology in the CFDS case, a T-S descriptor model of a heat exchanger given in [16] is considered. The effectiveness of the proposed fuzzy observer for the online estimation of unknown states of the used model is verified by numerical simulation. Figure 1: State variables x 1, x 2, x 4, x 5 and their estimates

Observer design for a class of T-S descriptor systems 2117 Figure 2: State variables x 3, x 6, x 7, x 8 and their estimates References [1] S. L. Campbell, Singular Systems of Differential Equations, Pitman, London, UK, 1980. [2] E. L. Lewis, A survey of linear singular systems, Circuits, Systems and Signal Processing, 5 (1986), no. 1, 3-36. http://dx.doi.org/10.1007/bf01600184 [3] L. Dai, Singular Control Systems, Springer, Germany, Berlin, 1989. http://dx.doi.org/10.1007/bfb0002475 [4] T. Takagi, M. Sugeno, Fuzzy identification of systems and its application to modeling and control, IEEE Trans. Syst., Man and Cyber., 15 (1985), no. 1, 116-132. http://dx.doi.org/10.1109/tsmc.1985.6313399 [5] T. Taniguchi, K. Tanaka, H. Ohtake and H. Wang, Model construction, rule reduction, and robust compensation for generalized form of Takagi-

2118 Abdellah Louzimi et al. Sugeno fuzzy systems, IEEE Transactions on Fuzzy Systems, 9 (2001), no. 4, 525-538. http://dx.doi.org/10.1109/91.940966 [6] K. Tanaka and H. O. Wang, Fuzzy Control Systems Design and Analysis: A Linear Matrix Inequality Approach, John Wiley & Sons, 2001. http://dx.doi.org/10.1002/0471224596 [7] Zs. Lendek, T. M. Guerra, R. Babuka and B. De Schutter, Stability Analysis and Nonlinear Observer Design using Takagi-Sugeno Fuzzy Models, Springer, 2011. http://dx.doi.org/10.1007/978-3-642-16776-8 [8] K. Tanaka, T. Ikeda, and H. O. Wang, Fuzzy Regulators and Fuzzy Observers: Relaxed Stability Conditions and LMI-Based Designs, IEEE Transactions on Fuzzy Systems, 6 (1998), no. 2, 250-265. http://dx.doi.org/10.1109/91.669023 [9] P. Bergsten and R. Palm, Thau-Luenberger observers for TS fuzzy systems, 9th IEEE International Conference on Fuzzy Systems, FUZZ IEEE 2000, San Antonio, TX, USA, (2000), 671-676. http://dx.doi.org/10.1109/fuzzy.2000.839073 [10] P. Bergsten, R. Palm and D. Driankov, Observers for Takagi-Sugeno fuzzy systems, IEEE Transactions on Systems, Man, and Cybernetics - Part B: Cybernetics, 32 (2002), no. 1, 114-121. http://dx.doi.org/10.1109/3477.979966 [11] D. Ichalal, B. Marx, J. Ragot, D. Maquin, Design of observers for Takagi- Sugeno discrete-time systems with unmeasurable premise variables, 5th Workshop on Advanced Control and Diagnosis, ACD, Grenoble, France (2007). [12] D. Ichalal, B. Marx, J. Ragot and D. Maquin, State estimation of nonlinear systems using multiple model approach, 2009 American Control Conference, (2009), 4636-4641. http://dx.doi.org/10.1109/acc.2009.5160393 [13] D. Ichalal, B. Marx, J. Ragot and D. Maquin, State and unknown input estimation for nonlinear systems described by Takagi-Sugeno models with unmeasurable premise variables, 17th Mediterranean Conference on Control and Automation, (2009), 24-26. http://dx.doi.org/10.1109/med.2009.5164542 [14] T. Taniguchi, K. Tanaka and H. O. Wang, Fuzzy Descriptor Systems and Nonlinear Model Following Control, IEEE Transactions on Fuzzy Systems, 8 (2000), no. 4, 442-452. http://dx.doi.org/10.1109/91.868950

Observer design for a class of T-S descriptor systems 2119 [15] Kilani Ilhem, Jabri Dalel, Saloua Bel Hadj Ali, Abdelkrim Mohamed Naceur, Observer Design for Takagi-Sugeno Descriptor System with Lipschitz Constraints, International Journal of Instrumentation and Control Systems (IJICS), 2 (2012), no. 2, 13-25. [16] J. Soulami, A. El Assoudi, M. Essabre, M. Habibi and E. El Yaagoubi, Observer Design for a Class of Class of Nonlinear Descriptor Systems: A Takagi-Sugeno Approach with Unmeasurable Premise Variables, Journal of Control Science and Engineering, 2015 (2015), Article ID 636909, 1-10. http://dx.doi.org/10.1155/2015/636909 [17] J. Soulami, A. El Assoudi, M. Essabre, E. El Yaagoubi, M. Habibi, Observer Design for a Class of Takagi-Sugeno Descriptor Systems: Application to an Aerobic Culture of a Recombinant Yeast, Applied Mathematical Sciences, 9 (2015), no. 61, 2999-3017. http://dx.doi.org/10.12988/ams.2015.5167 [18] Z. Wang, Y. Shen, X. Zhang and Q. Wang, Observer design for discretetime descriptor systems: An LMI approach, Systems & Control Letters, 61 (2012), no. 6, 683-687. http://dx.doi.org/10.1016/j.sysconle.2012.03.006 [19] V. Estrada-Manzo, Zs. Lendek, T. M. Guerra, Discrete-time Takagi- Sugeno descriptor models: observer design, 19th World Congress, IFAC, Cape Town, South Africa, 47 (2014), 7965-7969. http://dx.doi.org/10.3182/20140824-6-za-1003.00978 [20] B. Marx, D. Koenig and J. Ragot, Design of observers for Takagi-Sugeno descriptor systems with unknown inputs and application to fault diagnosis, IET Control Theory and Applications, 1 (2007), no. 5, 1487-1495. http://dx.doi.org/10.1049/iet-cta:20060412 [21] H. Hamdi, M. Rodrigues, Ch. Mechmech and N. Benhadj Braiek, Observer based Fault Tolerant Control for Takagi-Sugeno Nonlinear Descriptor Systems, International Conference on Control, Engineering & Information Technology (CEIT 13), Proceedings Engineering & Technology, 1 (2013), 52-57. [22] I. Hmaiddouch, B. Bentahra, A. El Assoudi, J. Soulami, E. El Yaagoubi, An Explicit Fuzzy Observer Design for a Class of Takagi-Sugeno Descriptor Systems, Contemporary Engineering Sciences, 7 (2014), no. 28, 1565-1578. http://dx.doi.org/10.12988/ces.2014.4858

2120 Abdellah Louzimi et al. [23] I. Hmaiddouch, B. Bentahra, A. El Assoudi, J. Soulami, E. El Yaagoubi, State Observer Design for a Class of Takagi-Sugeno Discrete-Time Descriptor Systems, Applied Mathematical Sciences, 9 (2015), no. 118, 5871-5885. http://dx.doi.org/10.12988/ams.2015.54288 [24] S. Boyd et al., Linear Matrix Inequalities in Systems and Control Theory, Society for Industrial and Applied Mathematics, Philadelphia, 1994. http://dx.doi.org/10.1137/1.9781611970777 Received: April 25, 2016; Published: June 24, 2016