Fuzzy model-based asynchronous H∞ filter design of discrete-time Markov jump systems☆
Introduction
In the control field, it is well known that the system states contain or reveal some essential information of dynamic systems, such as the voltage, the current, the position and the velocity in physical systems. However, state variables are not always measurable in control systems, in such case we have to estimate state variables for control purposes. Over the past decades, due to the great significance of state variables, state estimation or filtering problem has been one of the most important research topics and many effective techniques have been developed to deal with this problem, such as Kalman filtering [1], H∞ filtering [2], filtering [3], [4], [5]. Among all these techniques, H∞ filtering, which is also called energy-to-energy filtering, has been one of the most popular techniques. The essence of H∞ filtering is to minimize the gain between energy of the filtering error and that of the external disturbance [6]. Since H∞ filtering is even applicable for systems whose information of noise signals may be unknown, it has attracted considerable attention from researchers and a great deal of work related to H∞ filtering has been established [7], [8], [9].
In general, most of the dynamic systems suffer some random variations in parameters or structures in practice [10]. These random variations are usually caused by sudden environmental disturbance, component failures and component repairs [11], unfortunately, all of these have major influences on the performance of dynamic systems. To overcome this shortcoming, Markov jump systems (MJSs) have been introduced to describe dynamic models in various fields, such as biomedicine, economics and communication networks [12]. In the past decade, a number of results focused on H∞ filtering of Markov jump systems have been obtained. For example, H∞ filter design of Markov jump systems with sensor saturation and randomly occurring nonlinearities has been studied in [13], H∞ filtering for Markov jump systems with partly known transition probabilities has been reported in [14]. Resilient asynchronous H∞ filter design of Markov jump neural networks with multiplicative noises and unideal measurements has been considered in [15] while authors of [16] have studied H∞ filtering for Markov jump linear systems whose jumping mode information is not accessible.
It should be pointed out that most of the results aforementioned are concerned with linear Markov jump systems, however, we are more commonly faced with some nonlinearities in practical engineering problems, which increase the difficulties of the analysis and synthesis of control problems. Fortunately, the Takagi–Sugeno (T–S) fuzzy model bridges a gap between nonlinear systems and linear control due to its powerful approximation of any smooth nonlinear systems [17], [18], [19], [20]. Recently, T–S fuzzy model has been utilized successfully to deal with H∞ filtering problem of nonlinear Markov jump systems, for instance, H∞ filter for a class of nonlinear nonhomogeneous Markov jump fuzzy systems has been designed in [21], robust H∞ filtering for uncertain nonlinear singularly perturbed Markov jump fuzzy systems has been considered in [22], H∞ filter design for a class of nonlinear Markov jump neutral systems based on T–S fuzzy model has been studied [23]. Up to now, although so many results on H∞ filter design for Markov jump fuzzy systems have been reported, to the best of the authors’ knowledge, the issue of asynchronous H∞ filter design for discrete-time Markov jump fuzzy systems with uncertainties has not been fully addressed, this is the motivation of the present study.
In this paper, our main purpose is to solve the asynchronous H∞ filtering problem for a class of discrete-time Markov jump systems with uncertainties based on T–S fuzzy model, more specifically, we aim to design a filter which works asynchronously with the plant. Two different methods are proposed to deal with the filtering problem, which turn out to be equivalent. Sufficient conditions are derived to ensure the resultant filtering error system is stochastically stable with a prescribed H∞ performance level. The fuzzy-basis-dependent Lyapunov function is chosen to depend both on mode information of the plant and the filter, which is obviously less conservative than the common Lyapunov function approach due to all available information is fully utilized. To demonstrate the correctness and effectiveness of the proposed methods, a simulation example is provided.
The rest portion of this paper is organized as follows. In Section 2 the problem to be addressed is formulated while in Section 3 stochastic stability with a given H∞ performance of the filtering error system is analyzed. In Section 4, an sufficient condition is derived to deal with the filter design problem and the corresponding filter parameters are given explicitly. Finally, a simulation example is presented in Section 5 and the paper is concluded in Section 6.
Section snippets
Physical plant
Consider the following T–S fuzzy model based discrete-time Markov jump system with parameter uncertainties. Plant rule i: IF θ1k is Mi1, and and θdk is Mid, THENwhere Mij is a fuzzy set, and denote premise variables; and r represents the number of IF-THEN rules; xk ∈ Rl is the state; wk ∈ Rt is the disturbance that belongs to l2[0, ∞); yk ∈ Rg is the measurement output; zk ∈ Rv is the signal to be estimated;
Filtering performance analysis
In this section, the results of stability analysis and H∞ filtering performance analysis for the filtering error system (10) are carried out. The succeeding Theorem shows that the stochastic stability as well as a given H∞ performance of the filtering error system can be guaranteed if some sufficient conditions are satisfied and the filter parameters in Eq. (8) are known. Theorem 1 Consider the fuzzy system (5) and assume the filter parameters of system (8) are known. The filtering error system (10) is
H∞ filter design
After analyzing the stochastic stability and H∞ filtering performance of the filtering error system (10), now, based on the results of the obtained Theorems, we are in a position to design the asynchronous filter, that is, to decide the matrix parameters of the filter (8). Theorem 3 Given the fuzzy system (5), there exists a suitable filter in the form of Eq. (8) such that error system (10) is stochastically stable with a prescribed H∞ performance γ, provided we can find matrices
Illustrative example
In this section, we carry out a simulation example to show the effectiveness of the proposed filter design method.
Consider a tunnel diode circuit whose T–S fuzzy model was studied in [27]. The parameters of the T–S fuzzy model with uncertainties in the form of Eq. (5) are given as
Mode 1:Mode 2:
Conclusion
In this paper, we have addressed the problem of asynchronous H∞ filter design for a class of discrete-time Markov jump fuzzy systems with uncertainties. The designed filter can characterize asynchronous situation between the plant and the filter, which covers mode-independent or asynchronous filters as special cases. Two different but equivalent methods have been introduced to deal with the H∞ filter design synthesis, from which sufficient conditions are derived to guarantee the stochastic
References (27)
- et al.
Asynchronous l2–l∞ filtering for discrete-time stochastic Markov jump systems with randomly occurred sensor nonlinearities
Automatica
(2014) - et al.
Mode-dependent H∞ filtering for discrete-time Markovian jump linear systems with partly unknown transition probabilities
Automatica
(2009) - et al.
Design of H∞ filter for Markov jumping linear systems with non-accessible mode information
Automatica
(2008) - et al.
Robust H∞ fuzzy filter design for uncertain nonlinear singularly perturbed systems with Markovian jumps: an LMI approach
Inf. Sci.
(2007) A new approach to linear filtering and prediction problems
J. Basic Eng.
(1960)- et al.
Robust filtering for discrete-time Markovian jump delay systems
IEEE Signal Process. Lett.
(2004) - et al.
Energy-to-peak state estimation for Markov jump RNNs with time-varying delays via nonsynchronous filter with nonstationary mode transitions
IEEE Trans. Neural Netw. Learn. Syst.
(2015) - et al.
l2–l∞ filtering for multirate systems based on lifted models
Circuits Syst. Signal Process.
(2008) - et al.
On energy-to-peak filtering for nonuniformly sampled nonlinear systems: a Markovian jump system approach
IEEE Trans. Fuzzy Syst.
(2014) - et al.
On H∞ filtering for discrete-time Takagi–Sugeno fuzzy systems
IEEE Trans. Fuzzy Syst.
(2012)
A new design of delay-dependent robust H∞ filtering for discrete-time T–S fuzzy systems with time-varying delay
IEEE Trans. Fuzzy Syst.
Delay-dependent H∞ filter design for discrete-time fuzzy systems with time-varying delays
IEEE Trans. Fuzzy Syst.
A delay-dependent approach to robust H∞ filtering for uncertain distributed delay systems
IEEE Trans. Signal Process.
Cited by (34)
Stochastic stability analysis of nonlinear semi-Markov jump systems with time delays and incremental quadratic constraints
2023, Journal of the Franklin InstituteCitation Excerpt :The topic mainly includes stability analysis [5], controller design [6,7] and filter design [8]. From the Refs. [5–8], the sojourn time (ST) between two neighboring modes in MJSs obeys the exponential distribution (ED) or geometric distribution which satisfy memoryless property. Correspondingly, the transition rate (TR) between two different subsystems is a constant.
Asynchronous dissipative control design for semi-Markovian jump systems with uncertain probability distribution functions of sojourn-time
2021, Applied Mathematics and ComputationFinite-time asynchronous H<inf>∞</inf> filtering for positive Markov jump systems
2020, Journal of the Franklin InstituteCitation Excerpt :For example, the asynchrony resulting from overlapped detection delay is discussed in [16–19]. Asynchronous filtering is considered as Markov chain over a finite-time span in [20–23]. Wu et al. [24] is firstly concerned with the asynchronous issue for hidden Markov model.
Positive l<inf>1</inf>-gain asynchronous filter design of positive Markov jump systems
2020, Journal of the Franklin InstituteDiscrete-time fault tolerant control for semi-Markov jump systems with finite sojourn-time
2020, Journal of the Franklin InstituteCitation Excerpt :Markov chains is widely used in modeling failures of subsystems, sudden environment changes and electrical circuits with switches, such as in solar systems [1]. Memoryless property is the unique characteristic of a Markov chain, in which the transition rate between two modes is assumed to be constant and independent of the past states [2–6]. However, such characteristic is also a major limitation in modeling many real applications.
- ☆
This work is partially supported by Science Fund for Creative Research Groups of the National Natural Science Foundation of China (Grant no. 61621002), National Nature Science Foundation of China (No. 61633019), the Zhejiang Provincial Natural Science Foundation of China (No. LZ15F030004), Ningbo Science & Technology Planning Project (2014B82015) and the Australian Research Council (DP170102644).