- Research
- Open access
- Published:
The robust \(H_{\infty }\) control of stochastic neutral state delay systems
Journal of Electrical Systems and Information Technology volume 10, Article number: 39 (2023)
Abstract
In this paper, the resilient stochastic neutral state delay system with the Markov chain problem is examined. It offers output feedback control based on discrete state space delay and uncertainty that appear under traditional delay-independent conditions and exist under neutral stochastic state delays. Lyapunov theory can be used to solve the formulation \(x_c(t)=x(t)-Cx(t-\tau )\) for the intended \(H_\infty\) control of a stochastic model. Its stochastically mean square stability is shown using linear matrix inequalities (LMIs). The efficiency of the suggested strategy is shown by simulation results.
Introduction
Over the past few decades, neutral type stochastic state delay approach has become one of the most popular control strategies for a class of deterministic strict-feedback nonlinear systems. The neutral type schemes [2,3,4,5,6, 13, 16, 19] are all significant examples. Since stochastic neutral models are essential for many applications in engineering and research, numerous control mechanisms have been developed to ensure probability stability. In this study, robust stochastic state stability was established, and a neutral-type delay system was used in the design process together with \(H_\infty\) control.
For stochastic neutral delays, there are two different stability analyses: one is delay-dependent and contains information on delay sizes [3] through [9, 16, 17]; the other is delay-independent and can be applied to delays of any size [13]. Both employ LMI descriptions as part of their optimization techniques. In [10, 11], it is discussed how mixed nonlinear odd-even effects and arbitrary switching effects can be used to produce stochastic stability of finite duration. Time delays and parameter uncertainty are issues with the stochastic model’s \(H_\infty\) control; these issues were fixed when state feedback controllers could be constructed with the assumption that all state variables were present. The results cannot be applied if portions of the real states are unavailable. In [18] Introduced decomposition matrix technique with Jensen’s integral inequality, Peng-Park’s integral inequality, Leibniz-Newton formula and proved exponential stability of \(H_\infty\) performance level. In this paper, all the inequalities are exist in \(x(t)=f(t,x_t),\) do not given any importance for neutral delay. In [22] investigated finite-time bounded (FTB) tracking control for a class of neutral systems. Firstly, the dynamic equation of the tracking error signal is given based on the original neutral system. Then, it combines with the equations of the state vector to construct an error system, where the reference signal and the disturbance signal are fused in a new vector. The error system, input–output finite-time of the closed-loop system is proved stability by utilizing the Lyapunov–Krasovskii functional. The finite-time stability conditions are formulated in linear matrix inequalities (LMIs). It does not involved any neutral delay system.
The conventional delay-independent condition, which calls for a controller design with a uniformly sized time delay, is the only one that is used by the techniques. All physical systems cannot be applied because uniform time delay is not always present. To achieve less conservative delay-dependent conditions, many different approaches have been tried. An efficient stochastic differential system is determining the size of the delay under conservative assumptions of neutral type. The stability issue, however, clearly hinges on time delay because it is well known that a certain delay-independent criterion fails. It is required to show the time delay in the control design in order to produce delay-dependent conditions, which is necessary to get over this block [5, 7, 8, 12, 14, 15, 17,18,19, 21].
However, Chen et al. [2], Niculescu [19] studied the delay-dependent exponential stability of stochastic systems can similarly to x(t), which is stochastically stable in a deterministic manner. However, the presence of neutral stochastic case makes things more complicated and prevents the direct application of the techniques [6, 9, 17] of all deterministic state variables to a stochastic neutral system. Regardless of the parameter uncertainties, which have not yet been fully investigated, the stability criteria are dependent on dynamic output feedback controllers recommended by disturbance attenuation level; however, it is still an open and difficult problem. In this study, we demonstrated that the cap is stochastic by demonstrating that \(x_c(t)=x(t)-C x(t-\tau ), y(t)=f(t, x_t)\).
The paper is organized as follows. In “Preliminary modeling” section contains preliminary results. Some sufficient conditions are constructed for stochastic neutral system which is given in “Main models” section. In “The robust \(H_\infty\) control” section, the stability of state delay with \(H_{\infty }\) control of stochastic neutral system is given. In “Numerical examples” section presents two numerical examples to show the validity of the proposed results and some conclusions are drawn in “Conclusions” section.
Notations In this paper, \(R^n\) and \(R^{n\times m}\) denote, the n dimensional Euclidean space, respectively, and the set of all \(n\times m\) real matrices; \(L^2 [0, \infty )\) is the space of square integrable on the interval \([0,\infty ).\) The Euclidean norm \(|\cdot |\) is in \(R^n\) and \(C([-l,0]:R^n)\) which denote the family of continuous function \(\phi\) form \([-l,0]\) to \(R^n\) with the norm \(\Vert \phi \Vert =\sup _{-l\le \theta \le 0}|\phi (\theta )|,\) and I denote the compatible dimension identity matrix. The notation \(X\ge Y\) (respectively, \(X>Y)\) where X and Y are symmetric matrices, which is meant that \(X-Y\) is positive semi-definite (respectively, positive definite). The matrix A denotes the transpose of \(\lambda _{\text {max}}\) and \(\lambda _{\text {min}}\), it stands for eigenvalue of maximum and eigenvalue of minimum respectively and the operator norm is denoted by \(\Vert A\Vert =\sup \{|Ax|:|x|=1\}=\sqrt{\lambda _{\text {max}}(A^TA)}.\) The notation \((\Omega ,{{\mathcal {F}}},\{{{\mathcal {F}}}t\}_{t\ge 0},{{\mathcal {P}}})\) is complete probability space with a filtration \(\{{{\mathcal {F}}}\}_t\) and satisfying the usual conditions (i.e., the filtration contains all P-null set is right continuous). The space \(L^P_\mathcal{{F_0}}([-l,0]:R^n)\) denotes the family of all \(\mathcal {F_0}\)-measurable \(C([-l,0]:R^n)\)-valued random variables. The value \(\varphi =\{\varphi (\theta ):-l\le \theta \le 0\}\) is related to \(\sup _{-l\le \theta \le 0}E|\varphi (\theta )|^p<\infty .\)
Preliminary modeling
Let \(\{r(t),t\ge 0\}\) be a right continuous Markov chain on the probability space taking values in a finite state space \(S=\{1,2,\ldots ,N\},\) then following transition probabilities are holds [20];
where \(\Delta >0\) and \(\Lambda _{ij}\ge 0\) is the transition rate from i to j if \(i\ne j\) while \(\Lambda =-\sum _{j\ne i}\Lambda _{ij}.\)
If \(A^m=[a^m_{ij}]_{n\times n}\) and \(A^M=[a^M_{ij}]_{n\times n},\) satisfying \(a^m_{ij}\le a^M_{ij}, \forall 1\le i, j\le n,\) the interval matrix \([A^m,A^M]\) is defined by \([A^m,A^M]=\{A=[a_{ij}:a^m_{ij}\le a_{ij}\le a^M_{ij},1\le i,j\le n\}.\) If \(A,{\overline{A}}\in R^{n\times n},\) \(\Vert \Delta A\Vert \le {\overline{A}}\) is a nonnegative matrix, then \([A\pm {\overline{A}}]\) denotes the interval matrix \([A-{\overline{A}},A+{\overline{A}}].\) In fact, any interval matrix \([A^m,A^M]\) has a unique representation of the form \([A\pm {\overline{A}}],\) where \(A=(\frac{1}{2})(A^m+A^M),\) then \({\overline{A}}=(\frac{1}{2})(A^m-A^M).\)
For each \(i\in S\) and the interval matrices \([A_i\pm \overline{A_i}],[B_i\pm \overline{B_i}],[C_i\pm \overline{C_i}],[D_i\pm \overline{D_i}]\) are exists in S, then consider n-dimensional stochastic neutral state delay system with Markovian switching structure as
where \(x(t)\in R^n\) is the state vector, \(v(t)\in R^{n\times m}\) is the control input and \(w(t)\in R^q\) is an exogenous disturbance, \(z(t)\in R^p\) is an output controller, \(A,A_h,A_q,B_1,\) \(B_2,C,C_h,D_1,D_2\) are known real constant matrices with appropriate dimensions, \(q>0,\) \(h>0\) are constant time delays and q may not be equal to h, here \(l=\text {max}\{h,q\}\) and \(\phi (\theta )\) are continuously differentiable on the interval \([-l,0],\) \(\textrm{d}\omega (t)\) is a Brownian motion and \(E, E_h,H_1\) are stochastic constant matrices.
Main models
Consider state feedback controller is
to prove stability of system (2), the following assumptions are need. In this case the notation, A(r(t)) will be denoted as \(A_r\) and so on and time varying uncertainties are assumed as
where \(\Delta A_r, \Delta A_{qr}, \Delta B_{1r},\Delta B_{2r},\Delta E_{hr}, \Delta C_{r},\Delta H_{1r},\) \(\Delta H_{2r},\) \(\Delta D_{1r},\Delta D_{2r}\) are time varying uncertainty parameters. The resulting of the uncertain closed-loop system (2)–(3) can be written as
In this case, the controller access are depending on state response x(t) and the jumping process r(t). The solution of system (1)–(4) at time t is \(x(t,\phi ,r(t),v,w)\) which is related to the initial conditions \(\phi\) and jumping process r(t), and the control input v(t), disturbance w(t), respectively, and \(x(0,x_0,r(0),v,w)\) represents solution of system (4) at time \(t=0.\)
Definition 3.1
The state stochastic delay system (4) with \(q,h\in L^2([-l,0]:R^n),\) and for every \(\phi \in R^n\) exists, then solution \(x(t,\phi ,r(t),v,w)\) is said to be mean-square stochastically stabilizable if
and is said to be robustically stochastic stable with disturbance attenuation \(\gamma ,\) if it is mean-square stochastically stabilizable under zero initial conditions of \(E\Vert z\Vert _{L^2}\le \gamma \Vert w(t)\Vert _{L^2}\), then for all nonzero \(w\in L^2[0,\infty )\) such that
Let
and
and let \(G_{zw}\) denote the system exogenous input w(t) and control output z(t), then \(H_{\infty }\)-norm of the closed-loop system (1)–(4) is equivalent to \(\Vert G_{zw}\Vert <\gamma .\)
The following lemma provides some sufficient condition for stochastic stability of \(H_{\infty }\) control system.
Lemma 3.1
[20] Let \(Q_1\) be a nonnegative definite symmetric matrix, there exists an interval matrix \(A_r,\Delta A_r\in [A\pm {\overline{A}}]\) such that
and
Proof
It is well known that, if \(A_r=0\) or \(\Delta A_r=0,\) then [1]
If
for any \(\epsilon >0,\) then
And similarly if either \(A_r,E_r=0\) or \(\Delta A_r,\Delta E_r=0,\) then
\(\square\)
Theorem 3.1
Consider the neutral uncertain state delay system (4), if Markovian switching at \(v(t)=0\) and \(w(t)=0\) such that
There exist matrices \(P_i>0, Q>0,R>0,W>0\) and
such that system (4) is stochastically stable.
Proof
The Lyapunov functional of the above system is (see [8])
It can be derived by Ito’s formula;
where \(LV(x(s),s,r(s))={\dot{V}}(x(s),s,r(s)),\) L is the infinitesimal generator. Then
and
By Lemma 3.1 the above result can be rewritten as
where
and \(\eta (t)=[x(t),x(t-q),x(t-h)].\) Based on the inequality (9), mean-square stability of the system (3)–(4) can be proved if
where \(\Lambda =(x(t)-A_{q_i}x(t-q))^TP_i(x(t)-A_{q_i}x(t-q)) +\int _{t-q}^tx^T(\tau )Qx(\tau )\textrm{d}\tau +\int _{t-h}^tx^T(\tau )Rx(\tau )\textrm{d}\tau .\)
Note that, if \(\Xi _i<0,\) \(P_i>0\) and \(x\ne 0,\) such that
If
and \(\beta >0,\) then
Therefore by Ito’s formula
By Gronwell-Bellman lemma the above result can be rewritten as
for all \(Q>0,\) \(R>0.\) As on Lemma 3.1, the following inequalities are holds
and
for all \(r(0)\in S.\) Let
Taking the limit as \(T\rightarrow \infty ,\) for all \(T\in [0,\infty ),\)
If \(P_i>0,\) \(i\in S,\) then
where
this implies that closed-loop system (4) under the control law (3) is stochastically stable.
Form Schur complement, \(\Xi _i<0,i=1,\ldots ,s,\) if and only if
If \(X_i=P_i^{-1},\) then multiplying \(P_i^{-1}\) by (12), we get
As on LMI, the above expression can be rewritten as
where
\(\square\)
Theorem 3.2
The system (4) is stochastically stabilizable, if \(R>0,Q>0,W>0,X_i>0,P_i>0,\) \(Y_i,i=1,2,\cdots ,s\) exists and the disturbance \(w(t)=0,\) then controller \(K_i=Y_iX_i^{-1}\) is exists such that the following inequality holds
Proof
It follows that Theorem 3.1, \(\Xi _i\) can be written as
It is noting that above inequality and Theorem 3.1 plays an important role in proving stability of \(H^\infty\) control system. \(\square\)
The robust \(H_\infty\) control
Theorem 4.1
Given a scalar \(\gamma >0,\) \(Y_i,i=1,2,\ldots ,s\) and \(K_i=Y_iX_i^{-1}\) are exist in \(H_\infty\) control, there exist positive definite matrices \(P_i>0, R>0, Q>0,E_i>0, W>0\) such that system (1)–(3) is roust stochastically stable with given disturbance attenuation level \(\gamma ,\) if the following inequality holds
and
Proof
As on Theorem 3.1 the system (4) with \(q\in L^2([-l,0]:R^n)\) exists in \(H_{\infty },\) to prove \(H_{\infty }\) constraint of system (4) is mean-square stochastically stable, we need Lyapunov functional V(x(t), t, r(t)) and
It follows from Dynkin’s formula and fact that \(x(0)=0,\) then
Substitute (16) in (15), we get
Now let
Substitute (18) in (17), we get
where
and
If
and subsequently that \(E\Vert z(t)\Vert \le \gamma \Vert w(t)\Vert _{L^2},\) therefore closed-loop system (1–3) is robust stochastically stable. If Theorem 4.1 can be applied in the form of Schur complement, then \(J(t)<0,\) \(\Gamma _i<0,\) for all \(t>0.\) Let \(P_i=X_i^{-1}, Y_i=K_iX_i, T_i=\text {diag}(X_i,I,I,I,I),\) then multiplying (19) by \(T_i\) and \(\text {diag}(P_i^{-1},I,I,I,I)\), we can find the coupled matrix inequalities as follows
where
Hence, system (1)–(3) is roust stochastically stable with \(\gamma\) attenuation level. \(\square\)
Numerical examples
In this section, we give some numerical examples to demonstrate the effectiveness of the proposed results.
Example 1
Consider the neutral system (1)–(4) with \(K=Y_iX_i^{-1},\) and assume that following uncertainty matrix holds, the system (1) is stable if we choose state x(t) in \(R^3\) and Markov chain r(t) in \(S=\{1,2,3\}.\)
By Theorem 3.2, the closed-loop system (1–4) is stochastically mean square stable with respect of above uncertainty. The stability trajectory of (1)–(4) is shown in Fig. 1 (when \(\epsilon =1\)). This example shows that the value of \(x_1=x(t),\) \(x_2=x(t-q)\) and \(x_3=x(t-h)\) and we point that proposed algorithm result is more convenient in delayed sampled data and stochastic parameter systems. The responses of state feedback gain are
This objective is developed in MATLAB-LMI Control Toolbox.
Example 2
Consider the neutral system (1)–(4) with \(K=Y_iX_i^{-1},\) and assume that following uncertainty matrix holds, the system (1)–(4) is roust stochastically stable with \(\gamma\) attenuation level if we choose state x(t) in \(R^3\) and Markov chain r(t) in \(S=\{1,2,3\}.\)
By Theorem 4.1, the closed-loop system (1)–(4) is roust stochastically stable with \(\gamma\) attenuation level with respect of above uncertainty. The stability trajectory of (1)–(4) is shown in Fig. 2 (when \(\epsilon =0.25\)) and system responses feedback gain is showed in below and the convergence level of control design is showed in Table 1. This example shows that the value of \(x_1=x(t),\) \(x_2=x(t-q)\) and \(x_3=x(t-h)\) and we point that proposed algorithm result is more convenient in delayed sampled data and comparison is made some of the references at convergence level.
Remark 1
In example 1 if \(D_{2i}\) value is \(\left[ \begin{array}{cccc} 1 &{}0\\ 0 &{}0 \end{array}\right]\) and \(\gamma =1\) and in example 2 if \(D_{2i}\) value is \(\left[ \begin{array}{cccc} 0.2 &{}0\\ 0 &{}0 \end{array}\right]\) and \(\gamma =0.0517,\) then comparing these two example of \(K_i\), \(X_i\) and \(Y_i,\) values, only example 2 gives less conservative. If we choose more than (or) less than of 0.2 in \(D_{2i}\), the system trajectory is not vanishing. And also comparing reference [2, 22, 23], the existing result is more accurate. This example shows that our method could be lower attenuation level than existing result [2, 22, 23].
Conclusions
In this article, we looked into the Markov chain-based design of stochastic neural state delay systems. It offers predetermined uncertainties and stochastic neutral system’s discrete state space delay is stochastically mean square stable. In the \(H_\infty\) control design, the output feedback is used, whereas the LMIS filters design was employed to demonstrate robust stochastic stability. The design challenges were to turn a stochastic system’s state space delay into neutrality and solve a few reduced-order errors that converge to zero. Compared with the previous works, the main results of this paper have several features: (i) time delays can exist in control input and the measurement output, (ii) The uncertainty can appear in all system matrices. The proposed systems with the \(H_\infty\) control method can be achieved a lower attenuation level using the (iii) delay differential method [2]. The future research will be development of stabilization of some fractional delay systems of neutral type. The mathematical examples have shown significant improvements over some existing results.
Data availibility
Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.
References
Boyd S, Elghaoui L, Feran E, Balakrishnan V (1994) Linear matrix inequality in systems and control theory, vol 15. SIAM Studies in Applied Mathematics. SIAM
Chen WH, Zheng WX, Shen Y (2009) Delay-dependent stability and \(H_\infty\) control of uncertain neutral systems with time delay. IEEE Trans Autom Control 54:1660–1667
Chen H, Shi P, Lim CC, Hu P (2016) Exponential stability for neutral stochastic Markov systems with time-varying delay and its applications. IEEE Trans Cybern 46:1350–1362
Chen W, Xu S, Zhang B, Qi Z (2016) Stability and stabilization of neutral stochastic delay Markovian jump systems. IET Control Theory Appl 10:1798–1807
Doyle JC, Glover K, Khargonekar PP, Francis BA (1989) State space solutions to standard \(H_2\) and \(H_\infty\) control problems. IEEE Trans Autom Control 34:831–847
Fridman E (2001) New Lyapunov–Karsovskii functionals for stability of linear retarded and neutral type systems. Syst Control Lett 43:309–319
Fridman E, Shaked U (2002) A descriptor system approach to \(H_\infty\) control of linear time-delay systems. IEEE Trans Autom Control 47:253–270
Fridman E, Shaked U (2003) Delay-dependent stability and \(H_\infty\) control: constant and time varying delays. Int J Control 76:48–60
Han QL (2002) Robust stability of uncertain delay-differential systems of neutral type. Automatica 38:719–723
Huang S, Xiang Z (2016) Finite-time stabilization of switched stochastic nonlinear systems with mixed odd and even powers. Automatica 73:130–137
Huang S, Xiang Z (2016) Finite-time stabilization of a class of switched stochastic nonlinear systems under arbitrary switching. Int J Robust Nonlinear Control 26:2136–2152
Kuang Y (1993) Delay differential equations with applications in population dynamics. Academic, Boston
Mahmoud MS (2000) Robust \(H_\infty\) control of linear neutral systems. Automatica 36:757–764
Mahmoud MS, Xie L (1999) Stability and positive realness of time delay systems. J Math Anal Appl 239:7–19
Malisoff M, Mazenc F (2009) Constructions of strict Lyapunov functions. Communications and Control Engineering Series. Springer, London
Mazenc F, Malisoff M, Niculescu SI (2013) Stabilization of linear time varying systems with input delays: application to rapidly time varying systems. In: Proceedings of the 52th IEEE conference decision control, florence, Italy, pp 1584–1589
Mazenc F (2015) Stability analysis of time-varying neutral time-delay system. IEEE Trans Autom Control 60:540–546
Meesuptong B, Singkibud P, Srisilp P, Mukdasai K (2022) New delay-range dependent exponential stability criterion and \(H_\infty\) performance for neutral-type nonlinear system with mixed time-varying delays. AIMS Math 8(1):691–712
Niculescu SI (2001) On delay-dependent stability under model transformations of some neutral linear systems. Int J Control 74:609–617
Ravi Kumar R, Chong Kil To (2012) The stability of distributed neutral delay differential systems with Markovian switching. Chaos Solitons Fractals 45:416–425
Wu M, He Y, She JH (2004) New delay-dependent stability criteria and stabilizing method for neutral systems. IEEE Trans Autom Control 49:2266–2271
Wu J, Xu Y, Xie H, Zou Y (2023) Finite-time bounded tracking control for a class of neutral systems. Mathematics 11:1199–1215
Zhang X, Ruan D (2018) Exponential stability for neutral stochastic functional partial differential equations driven by Brownian motion and fractional Brownian motion. J Inequal Appl 2018:201–216
Funding
There is no funding available.
Author information
Authors and Affiliations
Contributions
R.R.K.: Conceptualization, Methodology, R.N.: Validation, Writing—original draft, V.A.: Writing—review & editing, A.S.: simulation development. All authors read and approved the final manuscript.
Corresponding author
Ethics declarations
Competing interests
The author declares that there is no competing interests in this paper.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Ravi Kumar, R., Naveen, R., Anandhi, V. et al. The robust \(H_{\infty }\) control of stochastic neutral state delay systems. Journal of Electrical Systems and Inf Technol 10, 39 (2023). https://doi.org/10.1186/s43067-023-00106-0
Received:
Accepted:
Published:
DOI: https://doi.org/10.1186/s43067-023-00106-0