PDC and Non-PDC fuzzy control with relaxed stability conditions for contintuous-time multiplicative noised fuzzy systems
Introduction
Based on the T–S fuzzy model approach [1], [2], nonlinear control systems have been received a great deal of attention over the last three decades. This is a successful and systematical approach for dealing with the stability of nonlinear control systems. Through the T–S fuzzy model, the trajectories of nonlinear system can be approximated by several linear subsystems with determined membership functions. This kind of model can provide an effective representation of complex nonlinear systems. In general, T–S type fuzzy model can be classified into homogenous T–S fuzzy model [3], [4], [5], [6] and affine T–S fuzzy model [7], [8], [9], [10]. Different from the homogenous T–S fuzzy model, the affine T–S fuzzy model has a nonzero bias on the linear subsystems. Through this reason, the affine T–S fuzzy models are difficult to study and analyze. Hence, the T–S fuzzy model considered in this paper is the homogenous one. In general, the PDC concept [3], [4], [5], [6], [7], [8], [9], [10], [11] is applied to design the fuzzy controller for T–S fuzzy models. According to PDC concept, the nonlinear controller can be blended by using linear feedback gains via fuzzy rules. Through the T–S fuzzy model and PDC design concept, the linear control criteria can be applied to analyze the complex nonlinear systems. However, a fuzzy control system with a large number of rules under PDC control concept requires higher computational demand solving the solution. Unlike the previous PDC-based fuzzy control [3], [4], [5], [6], [7], [8], [9], [10], [11], this paper intends to develop a Non-PDC fuzzy control scheme to eliminate the previous lack and to release the drawback of the conventional PDC stability conditions.
In the previously mentioned literature, many practical stochastic behaviors are inevitable and are described by random processes. In this situation, the systems to be controlled are always modeled by stochastic systems. Based on Itô's stochastic differential equation [12], the multiplicative noise term is used to represent the unpredictable stochastic behaviors of systems. The multiplicative noise term is structured as that states multiplied by the zero-mean white noise. In [6], [13], [14], [15], Itô's stochastic differential equation has been employed to represent nonlinear stochastic systems via T–S fuzzy model with multiplicative noise. Because of the above motivation, the control problem of a nonlinear continuous-time stochastic dynamic system modeled by the homogenous T–S fuzzy model with multiplicative noise is investigated in this paper.
Previously, most of the stability analysis bases on a single quadratic Lyapunov function, i.e.,V(x(t))=xT(t)Px(t) and P=PT>0. By the convex optimization algorithm, the stability and stabilization problems of fuzzy control can be converted into LMI problems [16]. Therefore, some transformation methods are used to derive the LMI stability conditions. The feasible solutions of the fuzzy controllers can be directly solved by using the convex optimization algorithm [16]. However, it is a hard task to find a common matrix P=PT>0 such that Lyapunov stability inequalities are all satisfied if the number of rules of a fuzzy system is large. In order to solve this problem, many efforts [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31] have focused on developing the relaxed stability conditions for T–S fuzzy models. In [17], [18], some relaxed quadratic stability conditions for fuzzy models were proposed. On the other hand, the nonquadratic Lyapunov function was also used to derive the relaxed stability conditions for the T–S fuzzy systems [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31]. This nonquadratic Lyapunov function is usually called as fuzzy Lyapunov function (FLF) [23], [24], [25], [26], [27] or parameter-dependent Lyapunov function (PDLF) [31]. The above relaxed quadratic and nonquadratic stabilization concepts are employed in this paper to develop novel relaxed stability conditions for the T–S fuzzy model with multiplicative noises. By solving these relaxed stability conditions, the stability of the closed-loop system can be achieved by the designed fuzzy controllers.
The relaxed stability conditions developed in [17] took into account the interactions among the fuzzy subsystems in an analytical manner by using the property of the quadratic form. Overcoming the conservatism of [17], the approach of [18] collects the interactions in a single matrix and represents it in terms of an LMI that can be solved in a numerical manner. Extending the approach of [18], a PDC-based fuzzy controller is designed in this paper by deriving the relaxed quadratic stability conditions for the T–S fuzzy models with multiplicative noises. In addition, a Non-PDC fuzzy controller design is also developed in this paper based on the nonquadratic Lyapunov function. Using the nonquadratic Lyapunov function, some scholars have derived the relaxed stability conditions for the discrete-time [19], [20], [21] and continuous-time [22], [23], [24], [25], [26], [27], [28], [29], [30], [31] T–S fuzzy systems. However, the multiplicative noise was not considered for the systems in [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31]. Considering the stochastic behaviors, a Non-PDC fuzzy control scheme is thus developed in this paper by using the nonquadratic Lyapunov function for the T–S fuzzy systems with multiplicative noises. Compared with the approaches of [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], the derivations of the proposed design scheme are more complex due to the difficulty in handling membership function's derivatives and dealing with the additional multiplicative noises, simultaneously. In this paper, the PDC fuzzy controller designed with relaxed quadratic stability conditions and the Non-PDC fuzzy controller designed with the relaxed nonquadratic stability conditions are both investigated. The proposed Non-PDC fuzzy controller is to provide an alternative solution to applications of which time derivatives of membership functions are available. In this paper, the stability conditions are directly derived of the LMI forms that can be used to solve stabilizing fuzzy controllers without using the completing square technique. In contrast to the PDC fuzzy controller, the Non-PDC fuzzy controller can provide a better methodology with regard to relax the conservativeness of stability and stabilization problems for the T–S fuzzy systems with multiplicative noises.
The organization of this paper is given as follows. Section 2 formulates the problem and presents preliminary results. Results of quadratic stability with PDC-based fuzzy control are discussed in Section 3. The Non-PDC fuzzy controller design via nonquadratic Lyapunov function is investigated in Section 4. In Section 5, two numerical examples are given to illustrate the effectiveness and applicability of the proposed design approaches. Finally, Section 6 provides some concluding remarks.
Notations: the symbol ⁎ in a symmetric matrix denotes the transposed element in the symmetric position. The E{Q(t)} denotes the expected value of Q(t). I denotes identity matrix with appropriate dimension.
Section snippets
Problem formulation and preliminaries
A fuzzy dynamic model has been proposed by Takagi and Sugeno [1] to represent local linear input/output relations of nonlinear systems. This fuzzy model is described by IF-THEN rules and has been employed to deal with the control design problems of nonlinear systems. Motivated by this, a general T–S fuzzy model is employed in this paper to represent a nonlinear stochastic system whose consequent parts are linear stochastic subsystems. The overall fuzzy model of the system is obtained by fuzzy
PDC-based fuzzy controller design via relaxed quadratic technique
Considering the stability analysis of classical T–S fuzzy model, some relaxed conditions have been proposed in [18]. In order to study the unpredictable stochastic behaviors of nonlinear dynamic systems, this section considers the multiplicative noises added in classical T–S fuzzy model as Eq. (2a), (2b). Therefore, one can obtain some novel results different to the approach developed in [18]. First, let us consider a Lyapunov function as follows:
According to the
Non-PDC fuzzy controller design via relaxed nonquadratic technique
In general, the quadratic Lyapunov function is usually used to derive the stability conditions for the PDC-based fuzzy controllers. However, the stability conditions developed based on the quadratic Lyapunov function remain more conservative characteristics [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31]. In order to obtain more relaxed stabilization conditions for continuous-time stochastic fuzzy dynamical models, a Non-PDC fuzzy controller is developed in this
Numerical examples
Example 1 Referring to [25], the following nonlinear system is considered to compare the applications of proposed design method with the approach developed in [14]: Considering the unpredictable stochastic behaviors of nonlinear system, the multiplicative noise is added into Eq. (32) and the system becomes Under x(t)∈[−d,d], β1=10 and β2=2.1, system (33) can be converted into the following T–S fuzzy model:
Conclusions
In this paper, a relaxed quadratic stability condition was derived to find the PDC-based fuzzy controllers for multiplicative noised T–S fuzzy systems. Overcoming the difficulty in handling membership function's derivatives and dealing with the additional multiplicative noises, a Non-PDC fuzzy controller was also developed based on the nonquadratic Lyapunov functions. It has been shown that the nonquadratic Lyapunov function is better than quadratic Lyapunov function in the derivations of
Acknowledgments
The authors wish to express their sincere gratitude to the anonymous reviewers who gave them some constructive comments, criticisms and suggestions.
References (32)
- et al.
Exponential stability analysis and controller design of fuzzy systems with time-delay
Journal of The Franklin Institute—Engineering and Applied Mathematics
(2011) - et al.
Constrained fuzzy controller design of discrete Takagi–Sugeno fuzzy models
Fuzzy Sets and Systems
(2003) - et al.
Robust fuzzy control of uncertain stochastic time-delay Takagi–Sugeno fuzzy models for achieving passivity
Fuzzy Sets and Systems
(2010) - et al.
Synthesis of nonlinear discrete control systems via time-delay affine Takagi–Sugeno fuzzy models
ISA Transactions
(2005) - et al.
Delay-dependent stabilization for stochastic fuzzy systems with time delays
Fuzzy Sets and Systems
(2007) - et al.
Passive fuzzy controller design for nonlinear systems with multiplicative noises
Journal of the Franklin Institute—Engineering and Applied Mathematics
(2010) - et al.
Stabilization via extended nonquadratic boundedness for constrained nonlinear systems in Takagi–Sugeno's form
Journal of The Franklin Institute—Engineering and Applied Mathematics
(2011) - et al.
LMI-based relaxed nonquadratic stabilizations for nonlinear systems in the Takagi–Sugeno's form
Automatica
(2004) - et al.
A new fuzzy Lyapunov function approach for a Takagi–Sugeno fuzzy control system design
Fuzzy Sets and Systems
(2006) - et al.
Exponential stability analysis and controller design of fuzzy systems with time-delay
Journal of The Franklin Institute—Engineering and Applied Mathematics
(2011)
Performance-oriented parallel distributed compensation
Journal of The Franklin Institute—Engineering and Applied Mathematics
Adaptive fuzzy systems and control: design and stability analysis
An approach to fuzzy control of nonlinear systems: stability and design issues
IEEE Transactions on Fuzzy Systems
Passive fuzzy controller design via observer feedback for stochastic Takagi–Sugeno fuzzy models with multiplicative noises
International Journal of Control, Automation, and Systems
Stability analysis and synthesis for an affine fuzzy control system via LMI and ILMI: continuous case
IEEE Transactions on Fuzzy Systems
Passive fuzzy control with relaxed conditions for discrete affine T–S fuzzy systems
International Journal of Innovative Computing, Information and Control
Cited by (40)
Dynamic behavior and sliding mode control on a stochastic epidemic model with alertness and distributed delay
2023, Communications in Nonlinear Science and Numerical SimulationRobust non-fragile proportional plus derivative state feedback control for a class of uncertain Takagi–Sugeno fuzzy singular systems
2019, Journal of the Franklin InstituteCitation Excerpt :The overall model of the nonlinear system is obtained by blending of these linear models through nonlinear fuzzy membership functions. Therefore, the control design is worked out by taking full advantages of linear control theory, and many efficient results on T–S fuzzy systems have been proposed (see, for example, [6–12], and the references therein). Recently, a wider class of fuzzy systems described by the singular form have been first considered in [13], which provided a new way to the analysis and synthesis of nonlinear singular systems.
Sliding mode fuzzy control for nonlinear stochastic systems subject to pole assignment and variance constraint
2018, Information SciencesCitation Excerpt :T-S fuzzy model is a powerful and efficient tool to handle complex nonlinear system and chaotic systems because a nonlinear system can be described as the ``blending'' of a set of linear dynamic models [32]. Many sufficient conditions for the stability of T-S model using a quadratic Lyapunov approach have been studied in [10,29,33,40,43]. However, various performance constraints have been considered besides the stability to obtain better performance, such as pole-placement constraints [24,38,44], covariance constraint [11], control input and output constraint [13,14,25,28–30].
Mean-square admissibility for stochastic T–S fuzzy singular systems based on extended quadratic Lyapunov function approach
2017, Fuzzy Sets and SystemsCitation Excerpt :Using the quadratic (non-quadratic) Lyapunov function, some scholars have derived the relaxed stability conditions for T–S fuzzy systems (see [22–33] and references therein), but the Brownian motions were not considered for the systems [19–33]. Compared with the approaches of [19–33], PDC and non-PDC fuzzy control design methods with relaxed stability conditions for continuous-time multiplicative noised fuzzy systems were proposed in [34]. However, to the best of our knowledge, the mean-square admissibility problem of stochastic T–S fuzzy singular systems has not yet been fully investigated.
New approaches to observer design and stability analysis for T–S fuzzy system with multiplicative noise
2017, Journal of the Franklin InstituteCitation Excerpt :Another important reason that caused conservativeness is the lacking of information of the membership functions in LMIs: most stability analysis methods are based on a single quadratic Lyapunov function in both fuzzy observer and controller design [27,28]. However, the region of attraction for fuzzy systems is MF (membership function) dependent [29–31], so it is a major challenge to find a single positive definite matrix to form a group of feasible LMIs for making the system stable [32], a class of fuzzy Lyapunov functions which are membership function dependent are proposed to solve this problem [25,33–35]. Recently, PMF (piecewise membership function) dependent method is proposed in [36–38] for common and fuzzy Lyapunov functions, but because of the involvement of membership functions derivatives, fuzzy Lyapunov functions cannot guarantee less conservativeness when it is compared with common Lyapunov functions in some specific stability regions [28], thus some additional matrices are used to relax the effects caused by derivatives of membership functions [39,40].