This paper proposes multicriteria adaptive observers for a class of singular systems with unknown time-varying parameters. Two criteria for the H∞ disturbance attenuation level and the upper bound of an ultimate invariant set are scalarized into a single cost function and then it is minimized by varying the weight parameter, which creates the optimal trade-off curve or Pareto optimal points. The proposed multicriteria adaptive observers are shown to be able to easily include integral action for better robust performance. It is demonstrated with numerical simulations that the proposed multicriteria adaptive observers provide the good estimation accuracy and allow effective and compromising design by considering two different cost functions simultaneously.
Ministry of Trade, Industry and Energy10062312Ministry of Science, ICT and Future Planning2016M1B3A1A01937655IITP-R0346-16-10071. Introduction
State estimation, or observation, has been recognized as one of the important research issues for dynamic feedback control systems since the full state information required for high performance is not available in most cases due to the high cost of sensors and limited accessibility for measurement. For state estimation, various types of observers have been developed, including Luenberger observers [1], sliding mode observers [2], and robust observers [3].
In the presence of unknown parameters encountered in most real systems, the observers designed for nominal models are hard to be applied in practical applications. For this reason, adaptive observers have been developed to estimate unknown parameters as well as state variables from input and output measurements, and hence achieve the robustness [4–8]. Recently, the results on adaptive observers have been successfully extended even to more general singular systems [9, 10]. Singular systems have extensive applications in many practical systems such as electrical systems, economics, mechanics, and chemical processes.
In implementing such practical adaptive observers over general singular systems, several criteria can be taken into account in consideration of design specifications. For example, adaptive observers can be designed according to the criteria such as H∞ [11, 12], H2, the ultimate region size, and so on. Mostly, among them, only one criterion has been employed for design of adaptive observers. However, two or more criteria could be applied to involve multiple design objectives, leading to a multicriteria optimization problem. Multicriteria based design enables us to do trade-off analysis for how much we must lose in one objective in order to do better in the other objective. For control design, the so called mixed criteria have already been adopted for practical implementation. As in control design, it would be meaningful to design adaptive observers with multiple useful criteria that can apply even to singular systems.
In this paper, we propose multicriteria adaptive observers for general singular systems with unknown time-varying parameters. For design of multicriteria adaptive observers, two criteria are employed to achieve robustness to disturbances and uncertainties. One is the H∞ attenuation level which is an upper bound on the H∞-norm of the transfer function from disturbances to estimation errors. The other is the upper bound of the ultimate region. These two criteria reflect how much disturbances and unknown parameters have effects on the estimation performance. Specially, the upper bound of the ultimate region makes the magnitudes of steady-state errors guaranteed to be upper bounded, which conflicts the H∞ criterion and hence provides an optimal trade-off curve and achievable values.
The optimal trade-off curve between the ultimate bound and the H∞ attenuation level is presented in the form of linear matrix inequalities (LMIs). Furthermore, the integrals of the error states are added for improving robustness to disturbances. If a singular matrix and time-varying parameters of the proposed multicriteria adaptive observers are set to be an identity matrix and constants, respectively, they reduce to existing adaptive observers for linear systems [13–15]. Simulation examples are presented to show the feasibility and the effectiveness of the proposed observers.
The paper is organized as follows: The description of multicriteria adaptive observers is given for a class of singular systems in Section 2. In Section 3, the design of multiobjective adaptive observers with integral effort is proposed. Finally, the simulation results are illustrated in Section 4 and the conclusion is drawn in Section 5.
2. Multicriteria Adaptive Observers
Let us consider the following singular system:(1)Ex˙t=Axt+Buut+Bθϕt,u,yθt+D1wt,yt=Cxt+D2wt,where x(t)∈Rn is the state, u(t)∈Rm is the input, θ(t)∈Rp is the unknown time-varying parameter, ϕ(t,u,y) is the nonlinear term depending on the input and the output, w(t) is the disturbance signal, y(t) is the measured output signal, and E, A, Bu, Bθ, C, D1, and D2 are the system matrices of appropriate dimensions. For a well-defined singular system, the rank of E is assumed to be r≤n. The nonlinear term ϕ(t,u,y) is known and upper bounded as(2)ϕt,u,y≤ϕmaxwith a certain positive constant ϕmax. In addition, it is assumed that uncertain parameters and their derivatives are upper bounded as(3)θ≤α,θ˙≤βwith positive constants α and β. Without loss of generality, the following conditions are also assumed to hold(4)rankEC=n,ranksE-AC=n,∀s∈C,where C is the set of complex numbers. Assumption (4) implies that the singular system (1) is observable. According to assumption (4), there exist nonsingular matrices T and N such that TE+NC=In, where In∈Rn×n denotes an identity matrix. The general solution for T and N is given as(5)TN=EC†+J1·I-ECEC†,where J1 is an arbitrary matrix of appropriate dimension and the superscript † denotes pseudoinverse. To estimate both the state variables and the unknown parameters, the following functional observer can be constructed:(6)z˙t=Fzt+Gut+Kyt+Bθ·ϕt,u,yθ^t,x^t=zt+Nyt,y^=Cx^t,where z(t) is the auxiliary variable of the observer, θ^ is the estimated parameter value, and F, G, and K are constant matrices to be determined later on for guaranteeing observation. It follows then that we have the following error dynamics:(7)e˙t=TA+FNC-KCxt-Fx^t+TBu-Gut+ND2y˙t+TBθϕt,u,yθt-Bθϕt,u,yθ^t+FND2-KD2+TD1wt,where e(t)=x(t)-x^(t). If F, G, and K in (7) are chosen to satisfy the following conditions:(8)TA+FNC-KC=F,TBu-G=0,ND2=0,the error dynamics (7) becomes(9)e˙t=Fet+Mϕθt+Bθϕeθt+FND2-KD2+TD1wt,where eθ(t)=θ(t)-θ^(t), M=(T-In)Bθ, and the arguments of ϕ(t,u,y) are omitted for simplicity. Substituting (8) into (9) yields(10)e˙t=TA+LpCet+Mϕθt+Bθϕeθt+D3wt,where Lp=FN-K and D3=LpD2+TD1. For the estimation of the unknown time-varying parameter θ(t), the following parameter update equation is constructed:(11)θ^˙t=ρaΓϕTUHCet+ΓUHCe˙t-Γρlθ^t,(12)e˙θt=θ˙t-θ^˙t=θ˙t-ρaΓϕTUHCet-ΓUHCe˙t+Γρlθ^t,where U,H are matrices to be designed, ρa is positive constant, Γ is a diagonal weight matrix for adaptation, and ρl is a leakage variable. The leakage term ρl is defined as [16](13)ρl=0if θ^<γthρa·ρ1if θ^≥γth,where γth≥α is a predefined threshold and ρ1 is a positive constant. In the estimation of unknown time-varying parameters, the function of last term in (12) is to force the estimated variable θ^ to inside of the set θ^<γth. Therefore, the term is effective when the estimated θ^ exists outside of the set. If the estimated parameter value θ^ exists outside of the set, ρaρ1 determines how fast it converges. Though there is a possibility of small oscillations on the switching surface θ^=γth, the leakage term ensures an bounded parameter estimation error. Also, these oscillations do not occur under nominal conditions. Furthermore, the parameter estimation is assumed to be independent of disturbances. Then, HD2=0 holds and the general solution is given as H=J2I-D2D2TD2-1D2T, where J2 is an arbitrary matrix. To derive an observer gain considering the effect of disturbance, the H∞ performance from the disturbance to the estimation error is defined as(14)supw∞≠0Ze∞w∞,where sup denotes supremum and Z is a matrix with appropriate dimension. Now, we shall try to construct sufficient conditions for multiobjective observer based on quadratic Lyapunov functions.
Theorem 1.
For given positive scalars τ, ρa, η1, η2, η3, and δ, if there exist matrices P=PT>0, Q=QT>0, R1=R1T>0, R2=R2T>0, R3=R3T>0, U, and X1 and scalars k1, k2, and k3 satisfying the following LMIs and the equality condition(15)minτ·ϵmax+1-τ·γ2(16)subject toR1PMϕmaxϕmaxTMTPTk1·I>0,R2Γ-1Γ-Tk2·I>0,(17)R3BθTPMϕmaxϕmaxTMTPTBθk3·I>0,(18)Ω=Ω11-1ρaBθPTA+X1CTPD3∗-2ρaBθTPBθϕmax+1ρaη2R2+1ρaη3R3-1ρaBθTPD3∗∗-γ2I<-δ·I,(19)BθTP=UHC,where(20)ϵmax=η1α2k1+η2ρaβ2k2+η3ρaα2k3,Ω11=ATTTP+PTA+CTX1T+XC+1η1R1+ZTZ,and ∗ denotes the entry of a symmetric matrix, then, the state estimation error and the parameter estimation error are uniformly ultimately bounded for an ultimate ellipsoidal set ϵmax with H∞ attenuation level γ. Moreover, the observer gain is chosen to be Lp=P-1X1.
Proof.
Choose the following Lyapunov function of a quadratic form:(21)Vet,eθt=eTtPet+1ρaeθTtΓ-1eθt.Differentiating the Lyapunov function (21) along the state trajectory yields(22)V˙t=2eTPFe+2eTPMϕθ+2eTPBθϕeθ+2eTPD3w+2ρaeθTΓ-1θ˙-2eθTϕTUHCe+2ρaeθTρlθ^-2ρaeθTUHCFe+Mϕθ+Bθϕeθ+D3w=2eTPFe+2eTPMϕθ+2eTPD3w+2ρaeθTΓ-1θ˙+2ρaeθTρlθ^-2ρaeθTBθTPFe+Mϕθ+Bθϕeθ+D3w,if BθTP=UHC is satisfied. The terms in (22) have upper bounds as follows:(23)2ρaeθTρlθ^≤2ρlρaθθ^-θ^θ^=2ρlρaγth-θ^θ^≤0,2eTPMϕθ≤1η1eTR1e+η1θTϕTMTPTR1-1PMϕθ,2ρaeθTΓ-1θ˙≤1ρaη2eθTR2eθ+η2ρaθ˙TΓ-TR2-1Γ-1θ˙,-2ρaeθTBθTPMϕθ≤1ρaη3eθTR3eθ+η3ρaθTϕTMTPTBθR3-1BθTPMϕθwhich comes from the following well-known inequality:(24)2xTy≤1ηxTSx+ηyTS-1y,S>0.Putting together inequalities in (23) and ignoring the effect of disturbances (i.e., w=0), we have(25)V˙t≤2eTPTA+LpCe-2ρaeθTBθTPTA+LpCe-2ρaeθTBθTPBθϕmaxeθ+1η1eTtR1et+1ρaη2eθTtR2eθt+1ρaη3eθTR3eθ+ϵ,where ϵ is defined by(26)ϵ=maxη1α2λmaxϕmaxTMTPTR1-1PMϕmax+η2ρaβ2λmaxΓ-TR2-1Γ-1+η3ρaα2λmaxϕmaxTMTPTBθR3-1BθTPMϕmax.The right hand side of inequality (25), except for ϵ, can be converted into an LMI and upper bounded as follows:(27)Ξ=Ξ1-1ρaBθPTA+X1CT∗-2ρaBθTPBθϕmax+1ρaη2R2+1ρaη3R3<-δ·I,where the Schur complement is used, PLp=X1, Ξ1=ATTTP+PTA+CTX1T+XC+1/η1R1, and δ is a design parameter to be chosen to be a small positive constant. If Ξ<-δ·I is satisfied, then, inequality (25) can be expressed as(28)V˙et<-δ·e+eθ+ϵ.It implies that V˙(e(t),eθ(t))<0 for δ·et+eθt>ϵ. When the estimation error exists outside of the bound, it approaches the inside of the bound and then it stays there according to the Lyapunov stability theory. Therefore, e(t), eθ(t) converge to the inside of a set parameterized by ϵ; that is, e,eθ∣e+eθ<ϵ/δ. It means the error dynamics is uniformly ultimately bounded with the ultimate bound ϵ/δ.
Now, the existence of disturbances is taken into account (the case of w≠0). For the H∞ performance supw2≠0Ze2/w2≤γ, the following inequality is considered with the derivative of Lyapunov function in (22).(29)V˙t+eTtZTtZtet-γ2wTtwt<0.Using Schur complement, (29) is equivalent to (18). From (28), the ultimate bound region ϵ is given as max[η1α2λmax(ϕmaxTMTPTR1-1PMϕmax)+η2/ρaβ2λmax(Γ-1R2-1Γ-1)+η3/ρaα2λmax(ϕmaxTMTPTBθR3-1BθTPMϕmax)]. Putting k1, k2, and k3 as an upper bound of each term yields(30)ϕmaxTMTPTR1-1PMϕmax≤k1·I,Γ-TR2-1Γ-1≤k2·I,ϕmaxTMTPTBθR3-1BθTPMϕmax≤k3·I.Applying the Schur complement, inequalities (30) are transformed to (16) to (17) in Theorem 1. Then, the maximum ultimate bound is given as ϵmax=η1α2k1+η2/ρaβ2k2+η3/ρaα2k3. Considering the maximum bound ϵmax and H∞ performance γ, multiobjective function can be constructed as(31)τ·ϵmax+1-τ·γ2,where 0≤τ≤1 is a weight parameter. This completes the proof.
3. Multicriteria Adaptive Observers with Integral Effort
In this section, the multiobjective adaptive observer involving integral action is presented to improve steady-state accuracy and attain the robustness to exogenous disturbances, which is of the following form:(32)z˙t=Fzt+Gut+Kyt+LIζ+Bθ·ϕt,u,yθ^t,ζ˙t=-Aζζ+yt-y^t,x^t=zt+Nyt,y^=Cx^t,where ζ is the integral of the estimation error. The proposed multicriteria adaptive observer (32) with integral effort yields the following error dynamics:(33)e˙=Fe-LIζ+Mϕθ+Bθϕeθ+LpD2+TD1w,ζ˙=-Aζζ+Ce+D2w.The following theorem tells us that the multicriteria adaptive observer (32) is guaranteed to achieve the H∞ attenuation level and the upper bound of the ultimate invariant set if some LMI conditions are met. ϵ¯max,γ¯ is used for the design of multicriteria with integral effort in order to distinguish them from ϵmax,γ for the one without integral term.
Theorem 2.
For given scalars 0≤τ≤1, η1>0, η2>0, η3>0, ρa>0, and δ>0, if there exist matrices P¯=P¯T>0, Q¯=Q¯T>0, R¯1=R¯1T, R¯2=R¯2T, R¯3=R¯3T, U¯, X¯1, X¯2, and Y¯ and scalars k¯1, k¯2, and k¯3 such that(34)minτϵ¯max+1-τγ¯2(35)subject toR¯1P¯MϕmaxϕmaxTMTP¯Tk¯1·I>0,R¯2Γ-1Γ-Tk¯2·I>0,(36)R¯3BθTP¯MϕmaxϕmaxTMTP¯TBθk¯3·I>0,(37)Υ1CTQ¯-X¯2-1ρaP¯TA+X¯1CTBθ+BθTP¯P¯TD1+X¯1D2∗-2Y¯-1ρaX¯2BθQ¯D2∗∗-2ρaBθTQ¯Bθϕmax+1ρaη2R¯2+1ρaη3R¯3-1ρaBθTP¯D3∗∗∗-γ¯2I<-δI,(38)BθTP¯=U¯HC,where(39)ϵ¯max=η1α2k¯1+η2ρaβ2k¯2+η3ρaα2k¯3,Υ1=ATTTP¯+P¯TA+CTX¯1T+X¯1C+1η1R¯1+ZTZ,then, the error dynamic (33) is ultimately bounded with an upper bound ϵ¯max and satisfies the H∞ performance with the γ¯ attenuation level. The observer gain is computed as Lp=P¯-1X¯1, LI=P¯-1X¯2, and Aζ=Q¯-1Y¯.
Proof.
Choose a Lyapunov candidate function as follows:(40)Vt=eTtP¯et+ζTtQ¯ζt+1ρaeθTtΓ-1eθt.Differentiating the Lyapunov function along the state trajectory with the condition BθTP¯=U¯HC results in(41)V˙=2eTP¯Fe-2eTP¯LIζ+2eTP¯Mϕθ+2eTP¯D3w-2ζTQ¯Aζζ+2ζTQ¯Ce+2ζTQ¯D2w+2ρaeθTΓ-1θ˙+2ρaeθTρlθ^-2ρaeθTBθTP¯Fe-LIζ+Mϕθ+Bθϕeθ+D3w,where equality BθTP¯=U¯HC is used. For now, the case of w=0 is considered.
Using (24) and Schur complement,(42)V˙=Υ+ϵ,where(43)Υ=Υ1CTQ¯-X¯2-1ρaP¯TA+X¯1CTBθ+BθTP¯∗-2Y¯-1ρaX¯2Bθ∗∗-2ρaBθTQ¯Bθϕmax+1ρaη2R¯2+1ρaη3R¯3,Υ1=ATTTP¯+P¯TA+CTX¯1T+X¯1C+1η1R¯1,X¯1=P¯Lp,X¯2=P¯LI,Y¯=Q¯Aζ,ϵ=maxη1α2λmaxMTP¯TR¯1-1P¯M+η2ρaβ2λmaxΓ-TR¯2-1Γ-1+η3ρaα2λmaxMTP¯TBθR¯3-1BθTP¯M.Then, the right hand side of inequality (42), except for ϵ, can be converted into (37) using the Schur complement and upper bounded. If Υ<-δ·I is satisfied, then, the inequality can be expressed as(44)V˙et<-δ·et2+ζt2+eθt2+ϵ.Therefore, e(t), ζ(t), and eθ(t) are uniformly ultimately bounded with the ultimate bound ϵ. The rest part is similar to that of Theorem 1, so it is omitted for brevity. This completes the proof.
Remark 3.
Since singular systems has a complicated structure, they provide more challenging issues. The proposed adaptive observer is more generalized than existing ones [7, 8, 13] that can be only applied to linear systems. Choosing T=I, N=0, it can be applied to standard linear systems. Further, the proposed one deals with time-varying parameters.
Remark 4.
The proposed observer offers flexibility between two criteria, an ultimate bound and H∞ performance, using a multiobjective approach. Until now, such design approaches for adaptive observers in singular systems with unknown time-varying parameters have not been studied at all.
4. Numerical Simulation
In this section, two examples for numerical simulations are considered to verify the effectiveness of the proposed multicriteria adaptive observers.
4.1. Example 1: A Second-Order Singular System
At the first example, the following second-order singular system is considered:(45)E=1000,A=-51-0.5-1,Bu=01,Bθ=01,C=1-1,D1=10,D2=0.The time-varying parameter is chosen to be θ=0.3sin(t) and ϕ is taken to be the sinusoidal function sin(5t). The parameters are chosen as ρa=10, η1=η2=η3=1, and δ=0.01. Applying Theorem 2, the optimal gains of the multiobjective proportional-integral adaptive observer with τ=0.1 are computed to be(46)LP=-10.2759-5.8751,LI=1.50870.4954,Az=0.6118.The solutions are provided with matrices(47)P¯=0.2059-0.1887-0.18870.1801,Q¯=0.3256,U¯=-0.1886,Y¯=0.1992,X¯1=-1.00770.8811,X¯2=0.2172-0.1955.In the presence of external disturbance w(t)=0.2sin(10t), the observer state tracks along a real state. By solving the multiobjective optimization problem, the optimal H∞ performance index is given as supw2≠0e~2/w2≤0.01839 and the optimal upper bound ϵ¯max=0.1855 is provided. Then, the system response curves of the system with the initial values x(0)=-2,2 are shown in Figure 1, which include the trajectories of state and estimated states. The oscillations in the estimation of states are caused by the external disturbances due to 0.2sin(10t) and nonlinearity ϕ(t,x,u). The parameter estimation curve is illustrated in Figure 2.
The state trajectories.
The time-varying parameter and its estimate.
4.2. Example 2: Leontief Model
In economics, the Leontief model describes the total production of the output required from each different industry to meet all demands. The model has been widely considered to predict the proper level of production of several types of goods. The state x represents the production of each industry, the matrix A corresponds to the rate of production, E is the stock placement of commodities, the input Buu(t) presents the known supply rate, θ is the external supply, the disturbance w(t) represents the uncertain industrial supply, and y corresponds to the production of commodities available for evaluation. For simulations, the system matrices are considered as follows:(48)E=1000001000001000000000000,A=120-621124-423211-520121-511121-5,Bu=10101,Bθ=00110,C=10100,D1=11000,with e(0)=10,20,20,17.5,17.5, ϕ(t,u,y)=sin(2t), θ(t)=sin(t), and w(t)=0.1cos(10t). The resulting LMI solutions given by Theorem 2 with τ=0.4 are(49)P-=0.6555-0.26090.12820.0892-0.1320-0.26090.29010.2008-0.15050.07130.12820.20080.4603-0.2110-0.00260.0892-0.1505-0.21100.29310.2255-0.13200.0713-0.00260.22550.5065,Q-=0.4192,U-=0.0574,Y-=0.2760,Aζ=0.6583,X-1=-0.9631-0.0849-1.0807-0.0330-0.1596,X-2=0.2269-0.00040.2238-0.0032-0.0130,Lp=5.59325.3555-11.5155-11.57225.4816,LI=-1.3456-1.30872.66882.7154-1.3873.The optimal value of multiobjective function is computed to be 0.8346. γ¯2=0.4773 and ϵ¯max=1.3706 are computed. Z=C is chosen for the H∞ performance index. In the presence of the disturbances, the estimation errors converge to zero. The convergence of error dynamics is presented in Figure 3. Figure 4 shows the trajectory of the parameter estimation errors. To show the trade-off between the ultimate bound ϵ¯max and the H∞ attenuation level γ¯, the optimal solutions are solved for various τ values. Plotting these optimal solutions, we obtain the Pareto optimal points as described in Figure 5. As shown in Figure 5, ϵ¯max seems to be inversely proportional to γ¯2. Figure 6 displays the comparison of the transient trajectories for different adaptive gains. To follow real parameter θ as fast as possible, a high adaptive gain is needed. However, if the adaptive gain is too large, it causes oscillations in the transient period but it has a good tracking performance for parameters as the case of ρa=20. Conversely, if the adaptive gain is too small, there is comparably small oscillations in the transient period but the observer provides a poor tracking performance for parameters as the case of ρa=5. Therefore, the adaptive gain should be appropriately chosen.
The estimation errors.
The parameter estimation.
The optimal trade-off curve for the multicriteria adaptive observer.
The comparison of parameter estimation with different adaptive gains.
5. Conclusion
Multicriteria adaptive observers were designed according to two criteria for the H∞ attenuation level of disturbances and the upper bound of the ultimate invariant set. The corresponding cost functions are scalarized into a single one and then the Pareto optimal solutions are obtained with Lyapunov stability in order to provide a good compromising solution. It was shown through numerical simulations that the proposed multicriteria adaptive observers have the good tracking ability.
For adaptive observers for general singular systems, other criteria can be easily taken into consideration by extending the proposed design scheme. It is believed that the proposed observers could be applied to fault detection, unknown input estimation, disturbance estimation, and so on.
Competing Interests
The authors declare that there are no competing interests regarding the publication of this paper.
Acknowledgments
This work was conducted under the framework of Industrial Technology Innovation Program of the Ministry of Trade, Industry & Energy (MOTIE, Korea) (10062312), Unmanned Vehicle Advanced Research Program funded by the Ministry of Science, ICT and Future Planning (MSIP, Korea) (2016M1B3A1A01937655), and ICT Consilience Creative Program through the IITP (Institute for Information & Communications Technology Promotion) supported by the MSIP (IITP-R0346-16-1007).
LuenbergerD. G.Observing the state of a linear system196482748010.1109/tme.1964.43231242-s2.0-84886166390SpurgeonS. K.Sliding mode observers: a survey200839875176410.1080/00207720701847638MR2422122ZBL1283.930662-s2.0-46149101014GuD.-W.PoonF. W.A robust state observer scheme200146121958196310.1109/9.975500MR1878221CarrollR. L.LindorffD. P.An adaptive observer for single-input single-output linear systems1973185428435MR0441420BesanG.Remarks on nonlinear adaptive observer design200041427128010.1016/s0167-6911(00)00065-7BesanconG.de Leon-MoralesJ.Huerta-GuevaraO.On adaptive observers for state affine systems200679658159110.1080/00207170600552766MR2216743VahedforoughE.ShafaiB.Design of proportional integral adaptive observersProceedings of the American Control Conference (ACC '08)June 2008Seattle, Wash, USA3683368810.1109/acc.2008.45870662-s2.0-52449087459PaesaD.FrancoC.LlorenteS.Lopez-NicolasG.SaguezC.Reset adaptive observer for a class of nonlinear systems201257250651110.1109/TAC.2011.21648192-s2.0-84856488095AlmaM.DarouachM.Adaptive observers design for a class of linear descriptor systems201450257858310.1016/j.automatica.2013.11.036MR31638072-s2.0-84893855247RodriguesM.HamdiH.TheilliolD.MechmecheC.BraiekN. B.Actuator fault estimation based adaptive polytopic observer for a class of LPV descriptor systems201525567368810.1002/rnc.3236MR33132972-s2.0-84922489446JungJ.HuhK.FathyH. K.SteinJ. L.Optimal robust adaptive observer design for a class of nonlinear systems via an H∞ approachProceedings of the American Control Conference2006Minneapolis, Minn, USA36383642AlmaM.AliH. S.DarouachM.GaoN.An H∞ adaptive observer design for linear descriptor systemsProceedings of the American Control Conference (ACC '15)July 2015Chicago, Ill, USA4838484310.1109/acc.2015.71720912-s2.0-84940906927JungJ.HwangJ.HuhK.Optimal proportional-integral adaptive observer design for a class of uncertain nonlinear systemsProceedings of the American Control Conference (ACC '07)July 2007New York, NY, USAIEEE1931193610.1109/acc.2007.42824962-s2.0-46449087276PaesaD.FrancoC.LlorenteS.Lopez-NicolasG.SaguesC.On robust PI adaptive observers for nonlinear uncertain systems with bounded disturbancesProceedings of the 18th Mediterranean Conference on Control and Automation (MED '10)June 2010Marrakech, Morocco1031103610.1109/med.2010.55476582-s2.0-77957009061PourgholiM.MajdV. J.A new non-fragile H∞ proportional-integral filtered-error adaptive observer for a class of nonlinear systems and its application to synchronous generators201122519911210.1243/09596518jsce1061IoannouP. A.SunJ.2012Chelmsford, Mass, USACourier