Elsevier

Systems & Control Letters

Volume 57, Issue 11, November 2008, Pages 904-912
Systems & Control Letters

Analysis and design of singular linear systems under actuator saturation and L2/L disturbances

https://doi.org/10.1016/j.sysconle.2008.04.004Get rights and content

Abstract

This paper carries out an analysis of the L2 gain and L performance for singular linear systems under actuator saturation. The notion of bounded state stability (BSS) with respect to the influence of L2 or L disturbances is introduced and conditions under which a system is bounded state stable are established in terms of linear matrix inequalities (LMIs). The disturbance tolerance capability of the system is then measured as the bound on the L2 or L norm of the disturbances under which the system remains bounded state stable and the disturbance rejection capability is measured by the restricted L2 gain from the disturbance to the system output or L norm of the system output. Based on the BSS conditions, assessment of the disturbance tolerance and rejection capabilities of the system under a given state feedback law is formulated and solved as LMI constrained optimization problems. By viewing the feedback gain as an additional variable, these optimization problems can be readily adapted for control design. Our analysis and design reduce to the existing results for regular linear systems in the degenerate case where the singular linear system reduces to a regular system, and to the existing results for singular systems in the absence of actuator saturation or when the disturbance is weak enough to not cause saturation.

Introduction

Singular linear systems have been drawing continual interest from control theorists for decades now (see, for example, [4], [8], [13], [15], [20] and the references therein). This is in part due to the many applications of singular systems in engineering, social and economic, and biological systems and in part due to the additional challenges that these systems present. For example, the algebraic constraints on system states that are inherent in singular systems lead to considerable complexities, such as impulses in system response.

The presence of actuator saturation imposes additional constraints on the analysis and design of singular systems. Addressing actuator saturation has been well recognized to be practically imperative yet theoretically challenging. This is reflected in the large body of literature on regular linear systems in the presence of actuator saturation (see, for example, [1], [11], [16], [29] and the references therein). This literature covers a wide range of control problems, including stabilization [17], [26], [28], [30], output regulation [5], [10], input output stability [2], [3], [18], the rejection of L2 and L disturbances [6], [7], [9], [23], [24], [25], [27], [31], [32], and stabilization in the presence of delays in the input [21], [22].

In comparison, analysis and design for singular linear systems subject to actuator saturation have been less studied. A few exceptions include [14], [19]. In particular, it is established in [14] that a singular linear system subject to actuator saturation can be semi-globally asymptotically stabilizable by linear state feedback if its reduced system under actuator saturation is semi-globally asymptotically stabilized by linear feedback. The latter was earlier established in [17], [16] to be possible under the assumptions that the open-loop system is semi-stable (i.e., with all its poles located in the closed left-half plane) and stabilizable in the usual linear system sense. More recently, we in [19] studied the stabilization problem for general, not necessarily semi-stable, singular linear systems by a saturated linear feedback law. A set of conditions were established under which the closed-loop system is regular and impulse free within an elliptical cylinder, and its cross-sectional ellipsoid is contractively invariant. With these set invariance conditions, which can be expressed in terms of linear matrix inequalities, the stabilization problem can be formulated and solved as an optimization problem with LMI constraints and the objective of maximizing the contractively invariant ellipsoid.

This paper continues from our earlier work [19] on the stabilization of singular linear systems under actuator saturation and carries out an analysis of the L2 gain and L performance for such systems. We will carry out such an analysis as follows. We will first introduce the notion of bounded state stability (BSS) with respect to the influence of L2 or L disturbances. We will then characterize the bounded state stability in terms of linear matrix inequalities (LMIs). We will measure the disturbance tolerance capability of the system by the bound on the L2 or L norms of the disturbances under which the system remains bounded state stable and the disturbance rejection capability by the restricted L2 gain from the disturbance to the system output or L norm of the system output. Based on the BSS conditions, assessment of the disturbance tolerance and rejection capabilities of the system under a given state feedback law is formulated and solved as LMI constrained optimization problems. By viewing the feedback gain as an additional variable, these optimization problems can be readily adapted for control design. Our analysis and design reduce to the existing results for regular linear systems in the degenerate case where the singular linear system reduces to a regular system [7], and to the existing results for singular systems in the absence of actuator saturation or when the disturbance is weak enough to not cause saturation [20].

More specifically, we consider the following singular linear systems subject to actuator saturation and disturbances, {Eẋ=Ax+Bsat(u)+Ww,u=Fx,z=Cx, where xRn is the state, uRm is the control input, wRr is the disturbance, zRp is the output, and sat:RmRm is a vector valued standard saturation function defined as sat(u)=[sat(u1)sat(u2)sat(um)]T,sat(ui)= sign (ui)min{|ui|,1}. Here we have slightly abused the notation by using sat to stand for both scalar valued and vector valued saturation functions. Also, we have assumed, without loss of generality, unity saturation level. Non-unity saturation levels can be absorbed into the matrices B and F. We also assume that the disturbance w belongs to one of the following two classes Wα2{w:R+Rr:0wT(t)w(t)dtα},Wα{w:R+Rr:wT(t)w(t)α,t0}, for some positive number α.

Our analysis is aimed to answer two questions. First, what is the maximal value of α such that, under the given feedback law u=Fx, the state remains bounded for all wWα2 (or wWα)? Such a question is related to the disturbance tolerance of the closed-loop system. When such a maximal α, say αmax, has been determined, our second question to answer is what the restricted L2 gain (or L norm of the output of the closed-loop system) is over a given Wα2 (or Wα) with an α(0,αmax]. This second question is related to the disturbance rejection capability of the closed-loop system. We will also be interested in the design of feedback gain F to increase the disturbance tolerance and rejection capabilities of the closed-loop system.

The remainder of the paper is organized as follows. Section 2 defines and characterizes the bounded state stability. Based on the characterization of bounded state stability, the disturbance tolerance and disturbance rejection capabilities are analyzed in Sections 3 Disturbance tolerance, 4 , respectively. Some numerical and simulation results are presented in Section 5 to illustrate the effectiveness of our analysis. Section 6 concludes the paper.

Section snippets

Bounded state stability

Consider a singular linear system under a saturated linear state feedback (1). Let rank (E)=q. Without loss of generality, assume that (E,A,B,W,C) are in the following form, E=[Iq000],A=[A11A12A21A22],B=[B1B2],W=[W1W2],C=[C1C2]. We will also partition the state vector accordingly as x=[x1x2],x1Rq,x2Rnq.

For the singular linear system under actuator saturation (1) with the matrices (E,A,B,W,C) in the form of (4), we are interested in the boundedness of its state trajectories. Such a system is

Disturbance tolerance

The disturbance tolerance capability of the closed-loop system under a given feedback gain F can be measured by the largest α, denoted as αF, such that the closed-loop trajectory under any disturbance wWα2 or wWα remains bounded. We will denote the maximal value of αF over the choice of F as α, i.e., α=supFαF.

L2 gain and L performance

The singular system subject to actuator saturation (1) is said to have a restricted L2 gain less than or equal to γ if, for x(0)=0 and any wWα2 compatible with x(0)=0, 0tzT(τ)z(τ)dτγ20twT(τ)w(τ)dτ,wWα2,t0. On the other hand, the output of the system has its L norm less than or equal to ζ if, for x(0)=0 and any wWα compatible with x(0)=0, zT(t)z(t)ζ2,t0.

The following theorem characterizes the conditions under which the system has a restricted L2 gain less than or equal to γ or the L

Example 1: L2 disturbance tolerance and rejection

Let us consider system (1) with E=[100010000],A=[0.60.80.50.80.600.610.8],B=[110.5],W=[0.10.10],F=[161],C=[111], and wWα2.

Let the set of initial conditions be specified by S=I2. Then, solving the optimization problem (22) leads to the following results: αF=51.3026, with H=[0.18381.65810],η=9.7266,and P1=103×[0.73720.31410.31411.6769],P3=103×[0.55770.0494],P4=0.3378×103. Shown in Fig. 1 is the ellipsoid E(P1,1+αFη) and a trajectory starting from a point on the boundary of the

Conclusions

This paper considered the problems of disturbance tolerance and rejection for singular linear systems subject to actuator saturation and L2/L disturbances. Under a given feedback, we developed methods for assessing the disturbance tolerance and rejection capability of the closed-loop system. Our analysis methods can be readily adapted for control design. Our analysis and design reduce to the existing results for regular linear systems in the degenerate case where the singular linear system

References (32)

  • L. Dai

    Singular Control System

    (1989)
  • R. De Santis et al.

    On the output regulation for linear systems in the presence of input saturation

    IEEE Transactions on Automatic Control

    (2001)
  • M. He et al.

    Structural decomposition and its properties of general multivariable linear singular systems

    Journal of Systems Science and Complexity

    (2007)
  • H. Hindi, S. Boyd, Analysis of linear systems with saturation using convex optimization, in: Proc. 37th IEEE Conf. Dec....
  • T. Hu et al.

    Output regulation of linear systems with bounded continuous feedback

    IEEE Transactions on Automatic Control

    (2004)
  • T. Hu et al.

    Control Systems with Actuator Saturation: Analysis and Design

    (2001)
  • Cited by (71)

    • Observer-based robust control for singular switched fractional order systems subject to actuator saturation

      2021, Applied Mathematics and Computation
      Citation Excerpt :

      In addition, actuator saturation often occurs in singular switched FOS and is often ignored [34–36]. If the controller designed parameters are beyond the working range of the actual systems, the performance of the closed-loop systems will be degraded [37–40]. Neverness, in reality,such as circuit systems, neural network systems, robot systems are suitable for this kind of systems, which also actively promotes the present work.

    • Finite-time dissipative control for singular discrete-time Markovian jump systems with actuator saturation and partly unknown transition rates

      2018, Applied Mathematical Modelling
      Citation Excerpt :

      And the observer-based H∞ control for uncertain singular time-delay systems with actuator saturation was investigated in [47]. Finite-time H∞ control for singular saturated switched system in [48], L2/L1 control for saturated singular systems in [49], control and and analysis for time delay saturation system in [50,51], observer-based stabilization for saturation system in [52] and so on. However, to the best knowledge of our authors, there are few results about finite-time dissipative control for discrete-time singular Markovian jump systems with actuator saturation and partly unknown transition probabilities.

    • Robust finite-time H<inf>∞</inf> control for discrete-time singular Markovian jump systems with time-varying delay and actuator saturation

      2016, Applied Mathematics and Computation
      Citation Excerpt :

      Thus, efforts were devoted to actuator and/or sensor saturation. For example, Ma et al. [37] studied singular T–S fuzzy systems under actuator saturation, Lin et al. [38] considered the robust control for systems with input saturation, the observer design problem in [39], singular switched time with actuator saturation in [40] and many other results in [41–43] and the references therein. In addition, time-delay, unavoidable in many practical systems, is also a major cause of instability and poor performance of dynamics systems.

    View all citing articles on Scopus

    Work supported in part by National Natural Science Foundation of China under the Overseas Young Investigator Program (Class B: Overseas Collaboration).

    View full text