A Nested Matrosov Theorem and Persistency of Excitation for Uniform Convergence in Stable Nonautonomous Systems

Size: px
Start display at page:

Download "A Nested Matrosov Theorem and Persistency of Excitation for Uniform Convergence in Stable Nonautonomous Systems"

Transcription

1 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 50, NO. 2, FEBRUARY A Nested Matrosov Theorem and Persistency of Excitation for Uniform Convergence in Stable Nonautonomous Systems Antonio Loría, Elena Panteley, Dobrivoje Popović, and Andrew R. Teel Abstract A new infinitesimal sufficient condition is given for uniform global asymptotic stability (UGAS) for time-varying nonlinear systems. It is used to show that a certain relaxed persistency of excitation condition, called uniform -persistency of excitation ( -PE), is sufficient for uniform global asymptotic stability in certain situations. -PE of the right-hand side of a time-varying differential equation is also shown to be necessary under a uniform Lipschitz condition. The infinitesimal sufficient condition for UGAS involves the inner products of the flow field with the gradients of a finite number of possibly sign-indefinite, locally Lipschitz Lyapunov-like functions. These inner products are supposed to be bounded by functions that have a certain nested, or triangular, negative semidefinite structure. This idea is reminiscent of a previous idea of Matrosov who supplemented a Lyapunov function having a negative semidefinite derivative with an additional function having a derivative that is definitely nonzero where the derivative of the Lyapunov function is zero. For this reason, we call the main result a nested Matrosov theorem. The utility of our results on stability analysis is illustrated through the well-known case-study of the nonholonomic integrator. Index Terms Matrosov theorem, nonholonomic systems, timevarying systems, uniform stability. I. INTRODUCTION IN MANY interesting nonlinear control problems, the closed-loop control system can be modeled by the timevarying, not necessarily periodic, differential equation When convergence of the trajectories of (1) to a given fixed point is required (this includes many trajectory tracking and adaptive control problems), perhaps the most appealing notion that includes such convergence is uniform asymptotic stability Manuscript received July 29, 2003; revised July 30, Recommended by Associate Editor Z.-P. Zhang. This work was supported in part by a CNRS-NSF collaboration project and was done in part while the first two authors were visiting the Center for Control Engineering and Computation, the University of California, Santa Barbara. This work was also supported by the Air Force Office of Scientific Research under Grants F and F and by the National Science Foundation under Grants ECS , ECS , and INT A. Loría and E. Panteley are with the C.N.R.S, UMR 8506, Laboratoire de Signaux et Systèmes Gif s/yvette, France ( loria@lss.supelec.fr; panteley@lss.supelec.fr). D. Popović is with United Technologies Research Center, East Hartford, CT USA. A. R. Teel is with the Center for Control Engineering and Computation, Department of Electrical and Computer Engineering, University of California, Santa Barbara, CA USA. Digital Object Identifier /TAC (1) (UAS) because of its inherent robustness (at least when has certain continuity properties that are uniform in ; see, e.g., [1, Lemma 5.4]) and its assertion that the convergence rate does not depend on the initial time in any significant way. See [2] for more detailed discussions on this topic and a list of related references. One can precisely characterize the property of UAS in many seemingly different, but in fact equivalent, ways. One way is via so-called estimates; that is, using a bounding function over the norm of the solutions that decreases uniformly with time and increases uniformly with the size of the initial states; see, e.g., [8]. Another common characterization is via the use of a smooth positive definite Lyapunov function with a uniformly negative definite total derivative; see, for instance, [1] and [3]. While these characterizations are very useful as intermediate steps in proving other properties (for instance, robust stability of a perturbed system), they can be difficult to establish. estimates are difficult to obtain because the solutions of (1) are not explicitly available. Lyapunov functions are appealing because explicit solutions to (1) are not needed. On the other hand, Lyapunov functions can be difficult to obtain because of the requirement that the derivative be uniformly negative definite. With regard to Lyapunov functions, in many model-based control applications it appears natural to use (when available) the closed-loop energy function as a candidate Lyapunov function. However, often the time derivative of this function is only negative semidefinite. For time-invariant problems, the typical analysis tool used to circumvent this problem is the Krasovskii La Salle invariance principle. This principle, the use of which requires some information about the solutions of the system, asserts that the trajectories converge to the largest invariant set contained in the set of points where the derivative is zero intersected with a level set of the Lyapunov function [4]. In the special case where this invariant set is the origin, asymptotic stability can be asserted [5]. The Krasovskii La Salle invariance principle is the key result that enables the so-called Jurdjević Quinn control algorithm for open-loop stable nonlinear control systems [6]. When the closed-loop is time-varying, one tool that is often used when the derivative of a Lyapunov function is only negative semidefinite is Barbălat s Lemma [7], [8]. This tool enables asserting that the derivative of the Lyapunov function converges to zero under certain regularity hypotheses. In adaptive control, Barbălat s Lemma is frequently relied upon to establish convergence to zero of part of the state. Barbălat s Lemma has also been used to establish convergence to the origin for a class of /$ IEEE

2 184 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 50, NO. 2, FEBRUARY 2005 nonholonomic systems controlled by smooth time-varying feedback. However, even when Barbălat s lemma can be used to show that the entire state converges to zero it typically cannot be used to conclude that the converge to zero is uniform in the starting time. It is well known that the invariance principle also applies to periodic time-varying systems (cf. [5] and [9]). Possibly motivated by these results, the study of stability for time-varying systems turned toward so-called asymptotically autonomous systems. Byproducts of this line of research are results based on the method of limiting equations. 1 Among the most significant results based on the method of limiting equations, we cite [12] where necessary and sufficient conditions for UAS are established. Along the same line of research is the recent work of [13] [16]. In these papers, the authors present a series of results among which we single out a generalization (in certain directions) of Krasovskii La Salle s invariance principle using the method of limiting equations. Roughly, in [16] UAS for a class of time-varying systems is concluded based on a reasoning a la La Salle on the system s equations considered as the initial time approaches infinity. The originality of this work resides in the introduction of different notions of uniform detectability which aid to conclude UAS for the case when one has a Lyapunov function with a negative semidefinite derivative. See [11] for an early reference on invariance principles for non autonomous systems using the formalism of limiting equations. For time-varying systems, another tool that has been used, but more sparingly, is Matrosov s theorem which first appeared in [17]; see also [18]. It pertains to the situation where one has a continuously differentiable Lyapunov function that establishes uniform stability and also a auxiliary function with appropriate properties. In particular, the auxiliary function should be bounded uniformly in time on bounded regions of the state space, and should have a definitely nonzero derivative on the set where a given, continuous, time-independent nonpositive upper bound on the Lyapunov function s derivative vanishes. Roughly speaking such property on the second auxiliary function allows to conclude that the trajectories cannot remain trapped in the set where the first function s derivative is zero but they necessarily converge to an invariant subset of the latter. Thus, Matrosov s theorem can be regarded, to some extent, as an invariance principle for nonautonomous systems. Significantly, it does not require any explicit information about the solutions of the system. It is purely an infinitesimal condition. While Matrosov s theorem relieves some of the burden from the first Lyapunov function, there are still no systematic methods for finding the second auxiliary function in Matrosov s theorem. A simplified version of Matrosov s theorem was used in [19] to establish one of the first results on uniform global asymptotic stability (UGAS) for robot manipulators in closed loop with a tracking controller. It also appears in the context of adaptive control in [20] and output feedback control, e.g., in [21]. There have been several interesting extensions of Matrosov s theorem over the years, mostly found in the work of [10], [18], and the references therein. In [22], a vector auxiliary function 1 Roughly speaking, dynamic equations describing the limiting behavior of the system when shifting time by where is an infinite unbounded sequence. For precise definitions and statements see [10, Chapter VIII, Section 5] and references therein as well as [11]. is used while in [23] and [24] a family of auxiliary functions is considered. It is worth emphasizing that in the latter reference the family of auxiliary functions is possibly uncountable and extensions are given that pertain to stability of sets. Moreover, locally Lipschitz Lyapunov and auxiliary functions are allowed. In all of these cases, the behavior of the auxiliary functions is referenced to the set where the upper bound on the derivative of the Lyapunov function vanishes. In [19], a simplified condition is given for checking that the auxiliary function has a sign-definite derivative. Most recently, in [25] the authors have extended Matrosov s theorem to pertain to differential inclusions, at the same time addressing stability of sets, using locally Lipschitz auxiliary functions and weakening the requirements on the upper bound of the derivative of the Lyapunov function. Despite all of the extensions of Matrosov s theorem, it is difficult to give general constructive methods for constructing Matrosov functions in the same way that it is difficult to give general constructive methods for finding Lyapunov functions notable recent exceptions are [26] and [63]. In order to meet the required conditions on the derivatives of Lyapunov (or auxiliary Matrosov ) functions, one approach is to use observability-type arguments and/or exciteness conditions. Roughly speaking one tries to identify a converging output and then verifies whether all the modes of the zero-dynamics (i.e., the dynamics which is left by zeroing the output) are sufficiently excited so that all converge to zero. For linear systems such methods have been under investigation, based on [27], starting probably with [28], and followed by numerous works including [29] [32]. In the nonlinear case we find for instance [2], [33], [34]. See also [15] which establishes sufficient conditions for uniform convergence in terms of a notion of (uniform) detectability and which is shown to be equivalent to different notions of persistency of excitation tailored for nonlinear systems and proposed therein as well as that presented in [2] (cf. Def. 4). The advantage of methods using observability-type arguments and related exciteness conditions is that for a certain class of systems one may infer the stability and convergence properties simply by looking at the dynamics of the system. For instance, for linear systems with bounded, it was shown in [30] that it is necessary and sufficient for uniform asymptotic stability that be persistently exciting (PE), namely that there exist and such that for all unitary vectors Equivalently, this system is uniformly (in ) completely (i.e., for all initial states) observable (see, e.g., [35]) from the output if and only if is PE. Notice that does not need to be full rank for any fixed. That PE is a necessary condition for uniform asymptotic (exponential) stability for linear systems has been well-known for many years now. In the case of general nonlinear systems, this was established for a generalized notion of persistency of excitation by Artstein in [12, Th. 6.2]. Relying on the notion of limiting equations it was proved in [12] that for the system (1) it (2)

3 LORÍA et al.: NESTEDMATROSOVTHEOREMANDPERSISTENCYOFEXCITATION 185 is necessary for uniform convergence that for each exist and such that there Sufficiency was also shown under other conditions involving Lyapunov functions and making use of a theorem establishing that uniform asymptotic stability of (1) is equivalent to asymptotic stability of all the limiting equations of (1). In view of the importance that persistency of excitation and related observability conditions have been proved to have in establishing convergence of linear and nonlinear systems, PE has been at the basis of the formulation of sufficient conditions for uniform asymptotic stability in many contexts. For example, see [12, Th. 6.3] and the results in [2], [15], and [27] [34] among others. In the context of adaptive control of nonlinearly parameterized systems, other notions of persistency of excitation for nonlinear systems have been introduced recently in [36]. In [2] and [37], we introduced a sufficient condition for uniform attractivity for a certain class of nonlinear systems. In words, our condition is that a certain function evaluated along the trajectories of the system, be persistently exciting whenever the trajectories are bounded away from a -neighborhood of the origin (cf. Def. 4). Loosely speaking, this property, called uniform -persistency of excitation ( -PE), ensures that the zero-dynamics of the system (with respect to a converging output) is sufficiently excited. We wrap up this brief review with [26] where the author constructs Lyapunov functions for time-varying systems based on Lyapunov functions with negative semidefinite derivative and by smartly exploiting the persistency of excitation of certain functions. See also [63]. A. Contributions of This Paper The contribution of this paper is twofold. On one hand, we present a new definition of -PE which we show to be necessary for uniform attractivity of the origin of general nonlinear time-varying systems. On the other hand, we present a generalized Matrosov theorem and show that one useful application of it is in proving that under some additional assumptions, -PE is sufficient for uniform attractivity. The new definition of -PE that we present here is stated in a form that does not involve the state trajectories (see Def. 3) and, therefore, it is simpler to verify than its predecessors (cf. [2] and [37]). As a matter of fact, it is also a reformulation of the condition (3) and for the system (1), we will show both necessity and sufficiency for uniform convergence without relying on limiting equations theory. The former follows simply using a Lipschitz condition and Gronwall s lemma. Our generalized Matrosov s theorem distinguishes itself from its predecessors in that it uses a finite family of auxiliary functions (as opposed to only one) to establish uniform convergence. For simplicity of exposition, we will limit our discussion to differential equations, uniform asymptotic stability of the origin, and a means of checking a sign-definiteness condition that is similar to what was used in [19]. Nevertheless, the results that we present here extend to more general settings (stability (3) of sets, differential inclusions, characterizations of signdefiniteness, etc.) Since we use a finite family of auxiliary functions, it is natural to wonder about the comparison to the vector auxiliary function used in [22] and the families of auxiliary functions used in [23] and [24]. The main difference is that, whereas in the previous references the behavior of all of the auxiliary functions and their derivatives is referenced to the set where the upper bound on the derivative of the Lyapunov function vanishes, our auxiliary functions are ordered and the behavior of an auxiliary function and its derivative is referenced to the set where the upper bounds on the derivatives of all of the preceding auxiliary functions vanish. We assume uniform stability, rather than assuming that we have a Lyapunov function which establishes it, and then none of our auxiliary functions is assumed to be sign definite. All of the auxiliary functions are assumed to be locally Lipschitz. We believe that the analysis tools that we present here may become an efficient tool to aid time-varying nonlinear control design, and may contribute to making the idea of using auxiliary functions for nonlinear systems, as introduced by Matrosov, a more versatile concept. A step forward in this direction is made in the early and extended versions of this paper [38], [39], where we solve some stabilization problems for fairly general nonlinear time-varying systems but having a clear impact in particular applications. See also [40] where necessary and sufficient conditions in terms of -PE and using our generalized Matrosov theorem, are established for the stabilization of interconnected driftless systems. This setting generalizes the popular benchmark of nonholonomic chain-form systems. The rest of the paper is organized as follows. In Section II, we define our notation and give some basic definitions. In Section III, we present our nested Matrosov theorem. In Section IV, we present the definition of uniform -PE ( -PE) and related properties. In Section V, we give a short proof of necessity of -PE and a result on sufficiency of -PE which covers Artstein s result. In Section VI, we illustrate, through the well-understood benchmark of nonholonomic systems, how to use our main results. We conclude with some remarks in Section VII. II. PRELIMINARIES Notation: Throughout this paper, stands for the Euclidean norm of vectors and induced norm of matrices, and, where, denotes the norm of time signals. In particular, for a measurable function, by we mean for and denotes the quantity. For two constants, we define. We also will use. A continuous function is of class if it is non decreasing. A continuous function is of class, if it is strictly increasing and ; if in addition, as. A continuous function is of class if for each fixed and as for each. Unless stated otherwise, throughout this paper we assume for the differential equation that is locally bounded,

4 186 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 50, NO. 2, FEBRUARY 2005 continuous almost everywhere, and that is continuous locally uniformly in. This property guarantees that solutions of exist locally in time, but they are not necessarily unique. We will use to denote a solution with initial condition. We will also use when we wish to leave the initial condition implicit. By abuse of notation, for a locally Lipschitz function and for all points where is differentiable, which is almost everywhere, we define. This abuse of notation is justified here as follows: Since is locally Lipschitz and each solution of is absolutely continuous, the usual time derivative exists for almost all time where the solution is defined. Moreover, with the properties we have assumed for and assuming the same properties for it follows from Fubini s theorem and the tools of nonsmooth analysis 2 that if for almost all then, for each solution we also have that for almost all. As it has been motivated, e.g., in [2] and [44], for time-varying systems the most desirable forms of stability are those which are uniform in the initial time. Definition 1 (Uniform Global Stability): The origin of the system (1) is said to be uniformly globally stable (UGS) if there exists such that, for each each solution satisfies Definition 2 (Uniform Global Attractivity): The origin of (1) is said to be uniformly globally attractive if for each, there exists such that Furthermore, we say that the origin of the system is uniformly globally asymptotically stable (UGAS) if it is UGS and uniformly globally attractive. This property is equivalent to the existence of such that, for all and A. Motivational Examples III. NESTED-MATROSOV THEOREM We present in this section our main theorem for analysis of time-varying systems. In the succeeding section we will present a fairly general result for UGAS of the origin of (1) based on the property of -PE and which makes use of our generalized Matrosov s theorem. To put our contribution in perspective we find it convenient to illustrate first how the original Matrosov s theorem (see [17], but also [18, Th. 5.5] and [8, Th. 55.3]) works on tracking control and time-varying stabilization problems. To that end, let us first look at the example of tracking control of robots as addressed in [19]. 2 For more precise statements, based on [41] and [42], see [25, proof of Prop. 1] and [43, item 5, p. 100]. (4) (5) (6) Example 1: Let us consider the Lagrangian model of a rigidjoints robot (see, e.g., [45]) where the inertia matrix is positive definite for all, is skew-symmetric and are control torques. The control problem is to make the robot follow a smooth reference trajectory such that. For this, we apply the following control law, originally proposed in [19] where,,. Now, we would like to analyze the stability of the closed-loop system, so we use the energybased Lyapunov function to obtain that and, hence, that the origin of the system is uniformly globally stable (UGS). From this equality, it is also fairly standard to invoke Barbălat s Lemma to conclude that (see, e.g., [46]). Alternatively, we can use the auxiliary function which is uniformly bounded on compact sets of the state and whose total time derivative satisfies 3 Intuitively, since we know that we may think that so, loosely speaking, if we wait long enough (and we can do that because the system is UGS) we will have that for large,. In particular, is square-integrable so invoking again Barbălat s Lemma one may accept that it should hold that as well. This idea can be made rigorous to actually prove UGAS via Matrosov s theorem: one needs to observe that is sign-definite (it is actually negative for all nonzero values of the state) on the set that is, on. The argument at work here is, roughly speaking, that the sign-definiteness of and the fact that on the set imply that the system s trajectories cannot remain trapped on the set unless they go to zero. See [8, p. 263] for a rigorous development on this idea and [19] for a rigorous analysis of this control system based on Matrosov s theorem. One important characteristic of the control problem shown previously that helps to use [18, Th. 5.5] is that the system is of relative degree one with respect to the output. Another is that the system is stabilizable by static feedback. However, for systems not satisfying these conditions, such as the nonholonomic chain of integrators, the analysis is more involved. In particular, one seemingly needs several auxiliary functions with appropriate properties in order to conclude attractivity of the origin. Yet, the intuition a la Krasovskii La Salle still holds as the following example illustrates. In Section VI, we establish precise 3 Using the facts that and. See, for instance, [45] for further details.

5 LORÍA et al.: NESTEDMATROSOVTHEOREMANDPERSISTENCYOFEXCITATION 187 results for nonholonomic systems using the Nested Matrosov theorem (cf. Theorem 1). Example 2: Consider the system in closed loop with 4 (7a) (7b) (7c) (8a) (8b) where and has certain excitation properties (to be specified). The closed-loop system becomes (with yet to be defined) (9a) (9b) (9c) The general intuition to establish a proof of asymptotic stability of the origin of (9) can be explained in terms of Krasovskii La Salle invariance principle. For this, let us restrict our attention to periodic feedbacks (as in [48]). First, taking the derivative of, we obtain that (10) From this inequality, we obtain that, hence, we may also admit from (9c) that. In addition to this, we have from (9b) that. This means that tends to a steady-state value which we denote by. Now, since is periodic in it is reasonable to assume that it is also sufficiently rich (persistently exciting) for each. If this is the case, then from the conjecture that necessarily the only constant value which may converge to is zero. Finally, the convergence of is obtained from (9a) if we define such that. The clear drawback of such an argument is that it cannot be made precise for general nonautonomous systems since it relies on Krasovskii La Salle invariance principle. However, with the aid of auxiliary functions one can establish the right convergence properties for each variable. Roughly speaking, similarly to the previous example, one needs to find an auxiliary function which allows to conclude that (also for not necessarily periodic functions ) on the manifold defined by. Then, one more function is needed to conclude that, under appropriate exciteness conditions,. The appropriate exciteness conditions may be thought of as a condition of positivity, in an averaged sense, of the function uniformly in ; this is formalized in Section III-B, in terms of U -PE. The Nested Matrosov theorem presented next formalizes the general intuition from the previous examples. The control problem of nonholonomic systems is addressed in Section VI in order to illustrate the use of use of Theorem 1. 4 This control law was proposed first in [47] see also [37] for a proof of UGAS and is used here for the sake of illustration only. B. Main Result Theorem 1 (Nested Matrosov Theorem): Under the following assumptions the origin of (1) is UGAS. 5 Assumption 1: The origin of the system (1) is UGS. Assumption 2: There exist integers, and for each there exist a number ; locally Lipschitz continuous functions, ; a function ; continuous functions, ; such that, for almost all, and all (11) (12) Assumption 3: For each integer, we have that 6 A), and all implies that B) for all. Assumption 4: We have that the statement A), and all implies that B). Theorem 1 generalizes, in certain directions, [25, Pro. 2] (see also [49, Prop. 2]) which, as clearly shown in that reference is, in turn, an extension of the classical Matrosov theorem [17] which combines an auxiliary function with a Lyapunov function that establishes UGS. See also the more recent expositions of Matrosov s theorem: [18, Th. 5.5, p. 58], [10, Th. 2.5, p. 62], and [8, Th. 55.3]. In particular, with respect to the formulation presented in [10, Th. 2.5, p. 62], it is worth remarking that: conditions i) and ii) imply UGS (assumed in Theorem 1), condition iii) implies the first part of (11) for the case of, condition v) (local uniform boundedness of ) is not imposed in Theorem 1 but we assume a local Lipschitz property. Besides the fact that the conclusion of Theorem 1 relies on finding, in general, more than one auxiliary function (besides the one to conclude UGS), the main difference with respect to [10, Th. 2.5, p. 62] is that we do not require that any of the functions to be sign-definite (see the cited reference for a precise definition) on the set where all the previous bounding functions, i.e., with, are zero. In particular, the bounds are allowed to depend on time through bounded (in ) functions as for instance in [19, Lemma in the Appendix]. Finally, we remark that Assumptions 3 and 4 imply Claim 1 (see further later) which in turn implies the sign-definiteness assumption in the original Matrosov s theorem. See [25, Prop. 2 and Cor. 1] for precise statements. 5 For clarity, we remind the reader that we do not assume that is locally Lipschitz in uniformly in as is common use. See Section II. 6 For the case this assumption takes the form: statement B) holds with.

6 188 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 50, NO. 2, FEBRUARY 2005 It can be useful to note that when the is locally Lipschitz uniformly in then the uniform global stability assumption can be relaxed to uniform (local) stability and boundedness of trajectories for each that is uniform in. In particular, the following hold. Theorem 2: If Assumption 1 in Theorem 1 is replaced by Assumption 5: 1) the origin is uniformly stable; 2) for each, there exists such that for all and ; 3) is locally Lipschitz uniformly in and Assumptions 2 4 hold then, the origin of (1) is UGAS. A sketch of proof is provided in the Appendix. C. Corollaries of Main Result A corollary of Theorem 2 generalizes well-known 7 results for time-varying nonlinear cascades (13) (14) Corollary 1: If, for (13) and (14), each initial condition produces trajectories that are bounded uniformly in the initial time, the functions and are locally Lipschitz uniformly in, and the origins of (14) and are UGAS then, the origin of (13) and (14) is UGAS. The following corollary, of Theorem 1, establishes that if one has a positive definite Lyapunov function with a negative semidefinite derivative bounded by a function that is observable, then UGAS follows. Observability is verified by differentiating the output as many times as needed so as to obtain a uniformly positive definite function, if possible. More precisely, consider (1) with a -times continuously differentiable output and define and for,. Define also. (15) Corollary 2: Consider the system (1) with output. Suppose there exist:, and a locally Lipschitz function such that and positive definite functions,, and such that (16) (17) (18) (19) Then, the origin is UGAS. Remark 1: In the autonomous case, i.e., when and are independent of time, (18) and (19) are implied by continuity of 7 The result that is best known is for the autonomous case and is due to [50], [51]. The corresponding result for the time-varying case, under the assumption of uniform boundedness of trajectories in both and, first appeared in [52]. and and positive definiteness of. In this case, the result of Corollary 2 is closely related to the results of [53, Sec. III]. Proof of Corollary 2: UGS holds by hypothesis [cf. (18) and (19)]. For the application of the nested Matrosov theorem, we take and, for, and. We obtain that and (20) (21) Observing that (19) implies the uniform boundedness, in, of for all, we have that there exists such that for all, and all. Then, we may invoke Theorem 2 with and the functions, for,, for and. Remark 2: Note that if we can write (22a) (22b) (22c).. (22d) (22e) with the bounded for bounded, given by (15) and redefined by (22), the result can be proved with the same functions. This extension is useful if and are not differentiable. In other words, Corollary 2 contains the special case where. D. Proof of the Main Result (Theorem 1) To prove the theorem we first need to establish the following claims. Claim 1: Given, there exists such that A) implies B). Proof: We prove the claim by contradiction. Suppose that for each integer, there exist such that for all, and. By compactness of, the continuity of, and Assumption 3, the sequence has an accumulation point such that for all. By Assumption 4, this implies that which contradicts the fact that. Claim 2: Let, and a continuous function be given. Then, Property 1 implies

7 LORÍA et al.: NESTEDMATROSOVTHEOREMANDPERSISTENCYOFEXCITATION 189 Property 2. Property 1: A) implies that B). Property 2: there exists A) such that According to the conditions of the theorem and the previous discussion, we have that, for almost all, and, using (11) together with (25) we obtain that for all (27) (28) implies that B). Proof: By Assumption 3 and, Property 2A implies that. Therefore, Property 2A implies (23) Now, if, then, due to Property 1, Property 2B holds for all whenever Property 2A holds. We claim further that there exists such that Property 2B holds whenever Property 2A holds and. Suppose not, i.e., for each integer there exists such that and (24) Then, by compactness of, continuity of, and Assumption 3, the sequence has an accumulation point such that. However, then from Property 1 we have that. By continuity of this contradicts (24) when is large and associated with a subsequence converging to the accumulation point. It now follows from the continuity of and compactness of that we can pick large enough to satisfy Using Assumption 1, for each and there exist and such that and Let and generate and and then let and generate and through the aforementioned claims and definitions. Let We claim that (29) (30) (31) (32) (33) Suppose not. It follows from the previous discussion that for all. Owing to the remarks in Section II on the derivative of locally Lipschitz functions along trajectories, it then follows that for almost all (34) Integrating and using (27), we have then Property 2A implies Property 2B. We now use these two claims to prove the theorem. According to Claim 1, Property 1 of Claim 2 holds when, and. An application of Claim 2 with these choices provides a value such that Property 1 of Claim 2 holds when, and. Continuing with this iteration, it follows that for each there exists and positive real numbers, such that, for all Next, define the locally Lipschitz function as (25) (26) which contradicts the choice of in (32). IV. UNIFORM -PERSISTENCY OF EXCITATION (35) In this section, we present a new definition of -PE, a property originally introduced in [2] and [37]. The newly defined property is conceptually similar to each of the previous ones; however, it is technically different in the sense that: first, it is easier to verify since it is formulated as a property inherent to a nonlinear function instead of being directly related to the solutions of a differential equation. Second, the new property is also necessary for uniform attractivity of the system (1). We also stress that, in general, neither -PE as defined in [2] nor as defined later, implies the other. Let be partitioned as where and. Define the column vector function and the set.

8 190 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 50, NO. 2, FEBRUARY 2005 Definition 3: A function where is locally integrable, is said to be uniformly -persistently exciting ( -PE) with respect to if for each there exist, and s.t. (36) If is -PE with respect to the whole state then we will simply say that is -PE. This notation allows us to establish some results for nonlinear systems with state by imposing on a certain function the condition of -PE w.r.t. only part of the state. For the sake of comparison, we recall next the definition proposed in [2] which is stated as a property of a pair of functions where is the vector field in (1) and the matrix function is such that is locally integrable for each solution of (1). Definition 4 [2]: The pair is called uniformly -persistently exciting ( -PE) with respect to [along the trajectories of (1)] if, for each and there exist constants and, s.t., all corresponding solutions satisfy for all. We emphasize that Definition 4 is cited here only for the sake of comparison. Throughout this paper, when we say that a function is -PE we mean in the sense of Definition 3. From now on, we will refer to the property defined in Definition 4 as -PE along trajectories. Even though in essence, the properties in both definitions are the same, as pointed out before, they are mathematically different. This is illustrated by the following example. Example 3: Consider the system A. Characterizations of -PE In this section, we present some useful properties which are equivalent to Definition 3. The proofs of all statements are omitted for space constraints (cf. [38]). Our first characterization of -PE applies to the particular (but fairly wide) class of uniformly continuous functions; we show that for such functions it is sufficient to verify the integral in (36) only for each fixed such that (i.e., for large states). Lemma 1: If is continuous uniformly in then is -PE with respect to if and only if A) for each there exist and such that, for all (38) The following Lemma helps us to see that Definition 4 states in words that a function is -PE with respect to if is PE in the usual sense 8 whenever the states (or similarly, the trajectories ) are large. This is important since it is the central idea to keep in mind when establishing sufficiency results based on the -PE property. This idea also establishes a relation with the original but also technically different definition given in [37]. It is also convenient to underline the similarity with the necessary condition (3), introduced in [12]. Lemma 2: The function is -PE w.r.t. if and only if B) for each and there exist and such that, for all (39) The last characterization is useful as a technical tool in the proof of convergence results as we show in Section V-B. Lemma 3: The function is -PE w.r.t. if and only if C) for each there exist and continuous strictly decreasing such that, for all (37) whose solutions with initial conditions and take the form,. Consider also the function Clearly, is locally integrable for each. One can see also that this function satisfies Definition 3 since but it does not satisfy Definition 4 with the initial conditions shown before, since. On the other hand, the function satisfies the trajectories-dependent property of Definition 4 but it does not satisfy Definition 3 for,. However, it is -PE in the sense of Definition 3 with respect to. (40) Remark 3: We may summarize the previous characterizations as follows. The following are equivalent: is -PE with respect to, statement B), and statement C). Also, each of these implies statement A. For uniformly continuous functions,, it is sufficient to check statement A), which on occasions may be easier to verify. Consequently, B) and C) will also hold. 8 That is, as defined for functions which depend only on time: that the function, is PE if there exist and such that for all unitary vectors we have that.

9 LORÍA et al.: NESTEDMATROSOVTHEOREMANDPERSISTENCYOFEXCITATION 191 Remark 4: A bibliographical remark seems adequate at this point. In [54] the authors pointed out that, for a nonlinear system where the time variations are due to an external input, the persistency of excitation as originally proposed for functions depending only on time, is neither sufficient nor necessary to assure uniform asymptotic stability. As a possible alternative, in [54, p. 157], the authors comment on the idea of defining a new persistency of excitation condition for the vector, with respect to the differential equation. This has been done in [2], [15], and [37]. In [12], a necessary condition, which is very close to -PE, for uniform (local) asymptotic stability is presented. It is important to mention that necessity for UAS is proved in that reference under more restrictive assumptions and in a less direct manner as we do in Section V-A. Most recently and independently, in [15] were reported several definitions of persistency-of-excitation for nonlinear time-varying systems. In particular, the author proposed the notion of output persistency of excitation (OPE) of the pair where is an output function and is as in (1). The interest of the results presented in [15] strives in different characterizations of a notion of uniform detectability previously introduced in [14] and which is shown to be necessary for UAS. It is also shown in [15] that detectability (in the sense defined in that reference) with respect to the output is equivalent to the property of -PE along trajectories, given in Definition 4. It can also be shown that Statement B in Lemma 2 is equivalent to OPE of the pair with respect to the set. See [15] for further details and definitions. B. Properties of -PE Functions We present here some important properties of -PE functions. Similar properties are well known for only-time dependent functions (cf. e.g., [32]) and some of which were proved to hold as well in [2], for pairs which satisfy Definition 4, i.e., along trajectories. The proofs are omitted due to space constraints and are provided in [38]. Let be -PE with respect to in the sense of Definition 3 and let,, 2 be continuous non decreasing functions. Assume that for all and almost all (41) (42) Property 1 (Multiplication of -PE Functions): If the function is -PE then, necessarily each is -PE. The opposite is not necessarily true. The proof of the first part of Property 1 follows directly using the Cauchy Schwartz inequality. A useful exception to the second part is the following. Property 2 (Power of a -PE Function): If the scalar function is -PE with respect to, with parameters, and then, for any the function is -PE with respect to, with parameters and,. It is also useful to remark that for functions linear in the state, -PE is equivalent to the usual PE property (when restricting to the set of nonnegative reals). Property 3: Consider the function where is locally integrable. Then, is -PE with respect to if and only if there exist and such that (43) The last property generalizes the well-known fact that for functions depending only on by which filtered (through strictly stable and proper transfer functions) regressors remain PE. This property is useful in control design. -PE Function): Consider the differ- Property 4 (Filtered ential equation (44) where is -PE with respect to and assume the following. 1) is such that 2) There exist, such that with denoting the solution of (44), we have that (45) (46) Then, defining and, the function is -PE with respect to. V. -PE IS NECESSARY AND SUFFICIENT FOR UGAS Making use of the tools previously introduced we show, for a fairly general class of nonlinear time-varying systems, that the property of -PE is necessary and sufficient for uniform attractivity of the origin. Sufficiency is established by a corollary of Theorem 1. A. Necessity The following result, contained in [12, Th. 6.2], gives conditions under which -PE of the right-hand side of a differential equation is necessary for uniform asymptotic stability. The technical conditions that we use permit a relatively straightforward proof, based on Gronwall s lemma, without recourse to the notion of limiting equations, as in [12, Th. 6.2]. Theorem 3 - : Assume that in (1) is Lipschitz in uniformly in. If the origin of (1) is UGAS, then is -PE with respect to. Proof: The Lipschitz assumption on implies that for each such that there exists such that (47)

10 192 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 50, NO. 2, FEBRUARY 2005 for all. From the UGAS assumption on (1) it follows that s.t. (48) Notice that without loss of generality, we may assume that for all. For the purposes of establishing a contradiction, assume that the statement of the theorem does not hold. More precisely, referring to the statement of Lemma 1, assume there exists such that for each and, there exist such that (49) Pick such that. Let the Lipschitz assumption generate the constant for and let. Let these values generate, and consider the solution of (1) starting at. By definition of solution which also satisfies (50) (51) In what follows, we use the following notation:,,, and correspondingly,. Theorem 4: Let Assumptions 1 3 hold. Suppose also the following. Assumption 6: We have that A) implies B). Assumption 7: The function, is independent of, locally Lipschitz in uniformly in, -PE w.r.t. and zero at the origin. Assumption 8: For all, we have where is continuous and. Then, the origin of (1) is UGAS. Remark 5: In Assumption 6, there is no requirement that the size of matches the size of. Proof of Theorem 4: This result follows from Theorem 1 by using the additional function We claim that this function satisfies for almost all, (54) Now, setting, we have from (49) that where (55) so using the Gronwall Bellman inequality, we obtain that On the other hand,, hence which contradicts (53). (52) (53) and represents a Lipschitz constant for on the set. Then, the result follows defining We now prove that (55) holds. Let the -PE property of and Lemma 3 generate the functions and. Then, for any, we have that B. Sufficiency This result derives from the sufficient conditions for UGAS established in Theorem 1. Accordingly, the following sufficient conditions are expressed in terms of a finite number of auxiliary functions (cf. Assumption 2) having a certain nested property (cf. Assumption 3) and the property that when the bounds on the derivatives of these auxiliary functions are all zero, this implies that part of the state and a certain function are zero (cf. Assumption 6) and, moreover, that the function is -PE with respect to the rest of the state (cf. Assumption 7). The last technical assumption (cf. Assumption 8) bounds the derivative of the rest of the state in terms of the part of the state and the function that are zero in Assumption 6. Second, the partial derivatives of are given by (56) The Lipschitz assumption on implies that for each there exists such that

11 LORÍA et al.: NESTEDMATROSOVTHEOREMANDPERSISTENCYOFEXCITATION 193 Hence, we obtain that and for almost all It follows that for almost all So, (55) follows using the bound from (56) in the previous inequality. It is worth pointing out that in (56) is not defined at [see statement C) in Lemma 3]. This motivates the definition of using the max of the two terms. The following corollary covers [12, Th. 6.3]. Corollary 3: If the origin of (1) is UGS and the following assumptions hold then, the origin of (1) is also UGAS. Assumption 9: For each, there exist a number ; a locally Lipschitz continuous function ; a continuous function ; such that, for almost all Assumption 10: We have that Assumption 11: We have that (57) (58) Assumption 12: The function is locally Lipschitz in uniformly in and -PE with respect to. VI. NONHOLONOMIC INTEGRATOR: A CASE STUDY Motivated by Brockett s celebrated paper [55], the problem of stabilization of nonholonomic systems of any dimension has been studied from numerous viewpoints. We suggest that interested readers see [56] for a tutorial with a very complete literature review up to Among recent works we cite [57], [58], and the seminal paper [59] which presents as a byproduct of the main results on controllability, universal controllers (i.e., for set-point and tracking) for practical stabilization of nonholonomic systems. In this regard, see also [60] where universal controllers achieving asymptotic stabilization for the (particular) case of underactuated (nonholonomic) ships are presented. The controllers used in this section are not original; we have sacrificed originality for clarity of exposition by choosing a well-understood benchmark to illustrate the use of the generalized Matrosov s Theorem 1 and the utility of the -PE property in control design. The controllers that we study were originally proposed in [47] where global asymptotic stability was shown and restudied in [61] where UGAS was established. We stress that as a byproduct of UGAS, one recovers the property of robust stability with respect to disturbances uniformly bounded in time, in the same spirit as for instance that of [62]. In this section, we provide a new direct proof of UGAS of the origin and furthermore, we establish necessary conditions for UGAS. Moreover, the results presented here can be generalized to a framework of interconnected driftless systems (which includes nonholonomic integrators) through time-varying nonlinear functions of the state. These results are not presented here for space constraints. Invited readers are referred to [39] and [40]. The approach to stability analysis presented below encompasses, in particular, the case of periodic time-varying feedbacks as considered for instance in [48]. With regard to this reference it is also interesting to observe that the author used Krasovskii La Salle invariance principle to obtain a direct proof of global asymptotic stability. Here, we use our generalization of Matrosov s theorem which may be considered as the extension of the invariance principle, for general nonautonomous systems. For clarity of exposition, we address separately the cases of three and more states. A. Case of Three States Consider the problem of stabilizing the nonholonomic chained system as described in Example 2. Then, we have the following result. Proposition 1: Consider the system (7) in closed loop with (8). Let the following hold. Assumption 13: The map be such that, all its first and second partial derivatives are uniformly bounded by where is a nondecreasing function 9 and, defining (59) assume that is -PE with respect to. Then, the origin of the closed-loop system is UGAS. Moreover, the origin is UGAS only if with is -PE with respect to. Remark 6: The necessary condition above is also sufficient. As a matter of fact, the -PE condition imposed on implies, via the filtering property (cf. Property 4), the -PE of with with respect to. B. Proof of Proposition 1 Necessity follows directly observing that UGAS implies, by Theorem 3, that is -PE, which implies in turn the statement of the proposition. The proof of sufficiency relies on Theorem 1 based on the intuition discussed in Example 2. First, UGS follows from the derivative of which yields (10). Integrating (10) from to we obtain that for all. Technically, this inequality is valid only on the interval of existence of the solutions. However, integrating on this window and using the fact that is bounded on the maximal interval of definition we proceed to integrate the -(9a) to obtain that 9 For simplicity, we use for a generic bound on any function which is uniformly bounded in.

12 194 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 50, NO. 2, FEBRUARY 2005 for all and where is a generic bound on which exists due to Assumption 13. Therefore, the solutions exist for all and actually, the origin is UGS. Notice also that here, since. To prove attractivity we consider other differentiable functions which are bounded and have bounded derivatives on balls of the state space. Our starting point in the pursuit of these additional functions to combine with is the observation that any terms in the derivative of subsequent auxiliary functions that vanish with, can be ignored since we know that. So, for example, we can take and, defining, we obtain that (60) From smoothness of and Assumption 13 we obtain that its total derivative is bounded for bounded, uniformly in. In the sequel, we will use the number as a generic bound on continuous functions over compact sets. With this under consideration, we have that For the sequel, we see that we can ignore terms in derivatives of auxiliary functions, that vanish with. As a matter of fact, from (7) and (8), we now see that we can ignore also terms that vanish with, and. What is more, if we were to consider the dynamics of the closed-loop system when and is constant the -equation would define the dynamics of a linear system with a time-varying input parameterized by a constant, i.e, Differentiating on both sides and owing to the fact that, we obtain that satisfies the differential equation where is defined in (59), and whose solution is (61) As we show farther down, is -PE. Based on these observations, we introduce the next function with the aim at concluding something about the difference between and the steady-state solution in other words, between and its steady-state solution. Hence, we define our third auxiliary function as, where along the trajec- and observe that the time derivative of tories of 10 (62) yields (63) To obtain this inequality, we have used the smoothness of all functions, Assumption 13 (which in particular implies that is locally Lipschitz, uniformly in ), and the compactness of. Our fourth function is introduced to be used in combination with the -PE property of in order to infer that the only constant value that may converge to is zero. This function is which satisfies for any (64) (65) and we claim that from the previous equation and the -PE assumption on, it follows that there exists a continuous, nondecreasing function such that and (66) This follows by appealing to Property 4 of -PE functions and observing that is defined by the differential equation Indeed, since is -PE with respect to so is. In other words, the steady-state value of when is also -PE. Then, appealing to Lemma 3 and using the inequality, the claim follows with 11. We proceed now to evaluate the time derivative of along the trajectories of the closed-loop system. To that end, we write (67) (68) (69) and observe that due to Assumption 13 all the partial derivatives in (67) and (68) which, are functions of, are uniformly bounded in by a generic bound that we denote by. Since also satisfies this boundedness property with we finally obtain, using (62), that for all and all 10 We emphasize that the choice of the function is inspired by the behavior of the system on the manifold, however, we consider its total derivative along the trajectories of the closed-loop system, on. In other words, we do not analyze the system s dynamics only on the defined manifold. 11 The use of the function stems from the fact that is not defined at zero.

13 LORÍA et al.: NESTEDMATROSOVTHEOREMANDPERSISTENCYOFEXCITATION 195 and using (66) and the fact that for all and we finally obtain that Thus, helps us to see that for the subsequent functions we can also ignore all the terms vanishing with. It is only left to find a function whose derivative is bounded by a negative term of and possibly positive terms involving, and. For this, we introduce whose total time derivative satisfies, for all and all Summarizing, we have that the functions satisfy where (70) (71) see (62) and (61) (72) for almost all (73) (74) C. Case of any Number of States We close Section VI with the extension of the previous result to the more general case of states, i.e., we now consider the set-point stabilization of the system (75a) (75b).. (75c) (75d) (75e) We use the following smooth control laws which are the counterparts for states of the controller (8): (76a) (76b) where for all and. That is, the occurrence of in alternates and the last term of is if is odd or, if is even. Interestingly, the last equations of the closed-loop system has the following form which reminds us of the controllability canonical form of linear systems. As shown first in [47] and also discussed in [61], one can write the equivalent closed-loop dynamics in the skew symmetric form (77a) where (77b) and,. So the result follows invoking the Nested Matrosov s Theorem 1. The verification of the assumptions is straightforward at this point. Remark 7: From a classical Lyapunov perspective, i.e., with aim at constructing a Lyapunov function with a negative definite derivative, the choice of is almost obvious from the structure of the last two closed-loop equations (cf. comments in Example 2). In particular, since we would be looking only for negative terms of and. The choice of comes almost naturally as a cross term that yields nonpositive terms of in. Notice that this would be an obvious choice if were constant (i.e., if the - subsystem were linear and strictly positive real). We also remark that, to some extent, this is the case of Example 1: a cross term of the type with and qualifies 12 as auxiliary function. Since is not constant nor positive for all and, we look for a function that guarantees that converges to its steady-state value while another function is used to exploit the -PE property via the characterization given by Lemma 3 to obtain a negative term of. The choice of the last function is rather obvious. 12 Notice that since is full rank, in the case of one degree of freedom we would have that necessarily for all which is tantamount to assuming constant in the current example (78) and are specific linear combinations of ; see [61]. Structurally, this system is a direct generalization of the closedloop equations for the case of three states. Notice in particular that one may easily conclude UGS of the origin. However, its analysis along the lines of [47] and [61] is much more involved than that of the previous case. In contrast to this, a direct proof of UGAS may be established via Theorem 1. To make a clear statement on the stability of (77), let us consider the system (1) with and, similar to (59), let us define Then, we have the following. (79). (80)

14 196 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 50, NO. 2, FEBRUARY 2005 Theorem 5: If the function defined above is -PE and the functions and are locally Lipschitz uniformly in and, the origin of (1), (79) is UGAS. If the origin is UGAS, then, with, is -PE with respect to. A direct consequence is the following. Proposition 2: Consider (75) in closed loop with (76). Let the following hold. Assumption 14: The map be such that, all its first and second partial derivatives be uniformly bounded by where is a nondecreasing function and let the function is -PE. Then, the closed-loop system (77) is UGAS. The proof of Theorem 5 is not presented here since it follows along the lines of the proof for the case of three states. However, for completeness we provide the guidelines for sufficiency. This follows by applying directly Theorem 1 with the following functions: (81) where with for all,. In particular, this function allows to show UGS. The rest of the auxiliary functions are with Defining for, and, finally (82) (83) (84) (85) Note the similarity with the functions for the case of three states. In particular, note that here we need functions as defined in (82) and functions as in (84) [see also (60) and (64)] where corresponds to the relative degree of (77b). VII. CONCLUSION In this paper, we have presented a new tool for establishing uniform attractivity of the origin when the origin is uniformly stable. The tool involves the use of an arbitrary finite number of auxiliary functions whose derivatives are simultaneously zero only at the origin. This result generalizes Matrosov s theorem on uniform asymptotic stability of the origin for nonlinear timevarying systems. We have also presented a new mathematical definition of the concept of -persistency of excitation which has been proved to be necessary and sufficient for UGAS of general nonlinear time-varying systems. Sufficiency is established via our generalized Matrosov s theorem. We have illustrated how to use our main results in the analysis of the popular benchmark of chained-form systems. APPENDIX SKETCH OF PROOF FOR THEOREM 2 The proof follows the same lines as the proof of Theorem 1. Up to (29), everything follows verbatim. From this point on, from Assumption 5. 2 we have that inequality (30) holds with and therefore,, and in (32) also depend on. Consequently, proceeding further as in the proof of Theorem 1, we obtain that for each and there exists such that (86) We use this fact together with ULS and the Lipschitz assumption to show uniform global boundedness. This together with ULS establishes UGS and so we recover all of the conditions of Theorem 1. We show uniform global boundedness by contradiction. Suppose that there exist a sequence of positive real numbers monotonically increasing to infinity,, and sequences with, and with such that (87) Let be an accumulation point for the sequence. Let generate according to Assumption In view of the ULS assumption, let be such that (88) Let and generate according to (86). Using Assumption 5.3, which implies continuity of solutions that is uniform on the intervals for all, we have that, for all sufficiently close to and all so that using the latter, (86) and (88), we get that This contradicts (87) for sufficiently large. REFERENCES (89) (90) [1] H. Khalil, Nonlinear Systems, 2nd ed. New York: Macmillan, [2] E. Panteley, A. Loría, and A. Teel, Relaxed persistency of excitation for uniform asymptotic stability, IEEE Trans. Autom. Control, vol. 46, no. 12, pp , Dec [3] J. Hale, Ordinary differential equations, in Interscience. New York: Wiley, [4] J. P. La Salle and S. Lefschetz, Stability by Lyapunov s Direct Method. New York: Academic, [5] N. N. Krasovskii, Problems of the Theory of Stability of Motion. Stanford, CA: Stanford Univ. Press, [6] V. Jurdjević and J. P. Quinn, Controllability and stability, J. Diff. Equat., vol. 28, pp , [7] I. Barbălat, Systèmes d équations différentielles d oscillations nonlinéaires, Revue de Mathématiques Pures et Appliquées, vol. 4, no. 2, pp , [8] W. Hahn, Stability of Motion. New York: Springer-Verlag, 1967.

15 LORÍA et al.: NESTEDMATROSOVTHEOREMANDPERSISTENCYOFEXCITATION 197 [9] J. P. La Salle, An Invariance Principle in the Theory of Stability, J. K. Hale and J. P. La Salle, Eds. New York: Academic, 1967, ch. Differential Equations and Dynamical Systems. [10] N. Rouche, P. Habets, and M. Laloy, Stability Theory by Liapunov s Direct Method. New York: Springer-Verlag, 1977, vol. 22. [11] Z. Artstein, Limiting Equations and Stability of Nonautonomous Ordinary Differential Equations, J. P. La Salle, Ed. Philadelphia, PA: SIAM, 1976, ch. Appendix A in Stability of Dynamical Systems. [12], Uniform asymptotic stability via the limiting equations, J. Diff. Equat., vol. 27, pp , [13] T. C. Lee, D.-C. Liaw, and B. S. Chen, A general invariance principle for nonlinear time-varying systems, IEEE Trans. Autom. Control, vol. 46, no. 12, pp , Dec [14] T. C. Lee and B. S. Chen, A general stability criterion for time-varying systems using a modified detectability condition, IEEE Trans. Autom. Control, vol. 47, no. 5, pp , May [15] T. C. Lee, On the equivalence relations of detectability and PE conditions with applications to stability analysis of time-varying systemss, in Proc. Amer. Control Conf., Jun. 2003, pp [16], Limit systems and attractivity for time-varying systems with applications to nonholonomic systems, presented at the 42nd. IEEE Conf. Decision Control, Maui, HI, [17] V. M. Matrosov, On the stability of motion, J. Appl. Math. Mech., vol. 26, pp , [18] N. Rouche and J. Mawhin, Ordinary Differential Equations II: Stability and Periodical Solutions. London, U.K.: Pitman, [19] B. Paden and R. Panja, Globally asymptotically stable controller for robot manipulators, Int. J. Control, vol. 47, pp , [20] R. Marino and P. Tomei, Global adaptive output feedback control of nonlinear systems. Part I: Linear parameterization, IEEE Trans. Autom. Control, vol. 38, no. 1, pp , Jan [21] S. Nicosia and P. Tomei, On the control of flexible joint robots by dynamic output feedback, in Proc. 4th. Symp. Robot Control, Capri, Italy, 1994, pp [22] N. Rouche, Attractivity of certain sets proved by using several Liapunov functions, in Symposia Matematica. New York: Academic, 1971, vol. 6, pp [23], On the stability of motion, Int. J. Nonlinear Mech., vol. 3, pp , [24] J. L. Corne and N. Rouche, Attractivity of closed sets proved by using a family of Lyapunov functions, J. Diff. Equat., vol. 13, pp , [25] A. Teel, E. Panteley, and A. Loría, Integral characterizations of uniform asymptotic and exponential stability with applications, Math. Control Signals Syst., vol. 15, pp , [26] F. Mazenc, Strict Lyapunov functions for time-varying systems, Automatica, vol. 39, pp , [27] R. E. Kalman, Contributions to the theory of optimal control, Bol. Soc. Mat. Mexicana, vol. 5, pp , [28] K. J. Åström and T. Bohlin, Numerical identification of linear dynamic systems from normal operating records, in Proc. 2nd IFAC Symp. Theory of Self-Adaptive Control Systems, P. H. Hammond, Ed., Teddington, U.K., 1965, pp [29] B. O. Anderson, Exponential stability of linear equations arising in adaptive identification, IEEE Trans. Autom. Control, vol. AC-22, no. 1, pp , Jan [30] A. P. Morgan and K. S. Narendra, On the uniform asymptotic stability of certain linear nonautonomous differential equations, SIAM J. Control Optim., vol. 15, no. 1, pp. 5 24, [31] S. Sastry and M. Bodson, Adaptive Control: Stability, Convergence and Robustness. Upper Saddle River, NJ: Prentice-Hall, [32] K. S. Narendra and A. M. Annaswamy, Stable Adaptive Systems. Upper Saddle River, NJ: Prentice-Hall, [33] D. Aeyels, R. Sepulchre, and J. Peuteman, Asymptotic stability for time-variant systems and observability: Uniform and nonuniform criteria, Math. Control Signals Syst., vol. 11, pp. 1 27, [34] Z. Artstein, Stability, observability and invariance, J. Diff. Equat., vol. 44, pp , [35] B. D. O. Anderson, R. Bitmead, C. Johnson Jr, P. Kokotović, R. Kosut, I. Mareels, L. Praly, and B. Riedle, Stability of Adaptive Systems. Cambridge, MA: MIT Press, [36] A. P. Loh, A. M. Annaswamy, and F. P. Skantze, Adaptation in the prescence of a general nonlinear parameterization: An error model approach, IEEE Trans. Autom. Control, vol. 44, no. 9, pp , Sep [37] A. Loría, E. Panteley, and A. Teel, A new persistency-of-excitation condition for UGAS of NLTV systems: Application to stabilization of nonholonomic systems, in Proc. 5th. Eur. Control Conf., 1999, Paper no [38] A. Loría, E. Panteley, D. Popović, and A. R. Teel. (2003) Persistency of excitation for uniform convergence in nonlinear control systems. Univ. California, Davis, CA. [Online]. Available: edu/math.oc/ [39], (2003) Matrosov s theorem using a family of auxiliary functions: An analysis tool to aid time-varying nonlinear control design. Univ. California, Davis, CA. [Online]. Available: edu/math.oc/ [40] A. Teel, A. Loría, E. Panteley, and D. Popović, Control of port-interconnected driftless time-varying systems, presented at the 7th. Eur. Control Conf., Cambridge, U.K., 2003, Paper 554. [41] F. H. Clarke, Y. S. Ledyaev, R. J. Stern, and P. R. Wolenski, Nonsmooth Analysis and Control Theory. New York: Springer-Verlag, [42] F. H. Clarke, Optimization and Nonsmooth Analysis. Philadelphia, PA: SIAM, [43] A. R. Teel and L. Praly, On assigning the derivative of a disturbance attenuation control Lyapunov function, Math. Control Signals Syst., vol. 13, pp , [44] A. P. Morgan and K. S. Narendra, On the stability of nonautonomous differential equations with skew-symmetric matrix, SIAM J. Control Optim., vol. 15, no. 1, pp , [45] M. Spong and M. Vidyasagar, Robot Dynamics and Control. New York: Wiley, [46] R. Ortega and M. Spong, Adaptive motion control of rigid robots: A tutorial, Automatica, vol. 25 6, pp , [47] C. Samson, Control of chained system: Application to path following and time-varying point stabilization of mobile robots, IEEE Trans. Autom. Control, vol. 40, no. 1, pp , Jan [48] Z. P. Jiang, Iterative design of time-varying stabilizers for multi-input systems in chained form, Syst. Control Lett., vol. 28, pp , [49] A. Teel, E. Panteley, and A. Loría, Integral characterizations of set UGAS/UGES with applications to Matrosov s theorem, presented at the IFAC NOLCOS, St. Petersburg, Russia, July 2001, Paper 214. [50] P. Seibert and R. Suárez, Global stabilization of nonlinear cascaded systems, Syst. Control Lett., vol. 14, pp , [51] E. D. Sontag, Remarks on stabilization and input-to-state stability, in Proc. 28th. IEEE Conf. Decision Control, Tampa, Fl, 1989, pp [52] E. Panteley and A. Loría, Growth rate conditions for stability of cascaded time-varying systems, Automatica, vol. 37, no. 3, pp , [53] C. Byrnes, A. Isidori, and J. C. Willems, Passivity, feedback equivalence, and the global stabilization of minimum phase nonlinear systems, IEEE Trans. Autom. Control, vol. 36, no. 11, pp , Nov [54] K. Narendra and A. Annaswamy, Persistent excitation in adaptive systems, Int. J. Control, vol. 45, no. 1, pp , [55] R. Brockett, Asymptotic stability and feedback stabilization, in Differential Geometric Control Theory, R. S. M. R. W. Brocket and H. J. Sussmann, Eds. Norwell, MA: Birkhäuser, 1983, pp [56] I. Kolmanovsky and H. McClamroch, Developments in nonholonomic control problems, Control Syst. Mag., pp , Dec [57] A. A. J. Lefeber, Tracking control of nonlinear mechanical systems, Ph.D. dissertation, Univ. Twente, Enschede, The Netherlands, [58] Z. P. Jiang and H. Nijmeijer, A recursive technique for tracking of nonholonomic systems, IEEE Trans. Autom. Control, vol. 44, no. 2, pp , Feb [59] P. Morin and C. Samson, A characterization of the Lie algebra rank condition by transverse periodic functions, SIAM J. Control Optim., vol. 40, no. 4, pp , [60] K. D. Do, Z. P. Jiang, and J. Pan, Universal controllers for stabilization and tracking of underactuated ships, Syst. Control Lett., vol. 47, pp , [61] A. Loría, E. Panteley, and K. Melhem, UGAS of skew-symmetric timevarying systems: Application to stabilization of chained form systems, Eur. J. Control, vol. 8, no. 1, pp , [62] C. Prieur and A. Astolfi, Robust stabilization of chained systems via hybrid control, IEEE Trans. Autom. Control, vol. 48, no. 10, pp , Oct [63] F. Mazenc and D. Nesic, Strong Lyapunov functions for systems satisfying the conditions of La Salle, IEEE Trans. Autom. Control, vol. 49, no. 6, pp , Jun

16 198 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 50, NO. 2, FEBRUARY 2005 Antonio Loría was born in Mexico in He received the B.Sc. degree in electronic engineering from the ITESM, Monterrey, Mexico, in 1991, and the M.Sc. and Ph.D. degrees in control engineering from the UTC, France, in 1993 and 1996, respectively. From December 1996 to Dec. 1998, he was successively an Associate Researcher at the University of Twente, Enschede, The Netherlands; NTNU, Norway, and the CCEC of the University of California, Santa Barbara. He is currently Charge de Recherche at the French National Centre of Scientific Research (CNRS). He is an author/coauthor of more than 80 scientific articles and the books Passivity Based Control of Euler Lagrange Systems ( New York: Springer Verlag, 1998) and Control of Robot Manipulators in Joint Space (New York: Springer Verlag, 2005). His research interests include modeling and control of Euler Lagrange systems, stability analysis of nonlinear time-varying systems, biped locomotion, and output feedback stabilization. He is an Associate Editor of Systems and Control Letters. Detailed information and publications are available at Elena Panteley was born in Leningrad, U.S.S.R. She received the M.Sc. and Ph.D. degrees in applied mathematics from the State University of St. Petersburg, Russia. She holds a research position with the French National Centre of Scientific Research (CNRS), at Laboratoire de Signaux et Systèmes. From 1986 to 1998, she held a research position with the Institute for Problem of Mechanical Engineering of the Academy of Science of Russia, St. Petersburg. During 1998, she was an Associate Researcher at the Center for Control Engineering and Computation of the University of California, Santa Barbara. During 1999, she was with the INRIA Rhône Alpes, Monbonnot, France. She is a coauthor of more than 70 scientific articles and book chapters. Her research interests are stability of nonlinear time-varying systems, control of electromecanical systems, and nonlinear and robust control. Dobrivoje Popović received the B.S. degree in electrical engineering from the University of Belgrade, Belgrade, Serbia, in 1998, and the M.S. and Ph.D. degrees in electrical and computer engineering from the University of California, Santa Barbara, in 2000 and 2004, respectively. He is currently a Senior Engineer at the United Technologies Research Center, East Hartford, CT. His research interests include extremum seeking, nonlinear optimization, and multivariable control. He is a Member of the Society of Automotive Engineers. Andrew R. Teel received his A.B. degree in engineering sciences from Dartmouth College, Hanover, NH, in 1987, and the M.S. and Ph.D. degrees in electrical engineering from the University of California, Berkeley, in 1989 and 1992, respectively. After receiving the Ph.D., he was a Postdoctoral Fellow at the Ecole des Mines de Paris, Fontainebleau, France. In September 1992, he joined the Faculty of the Electrical Engineering Department at the University of Minnesota, Minneapolis, where he was an Assistant Professor until September of In 1997, he joined the faculty of the Electrical and Computer Engineering Department at the University of California, Santa Barbara (UCSB), where he is currently a Professor. He is currently the Director of the Center for Control Engineering and Computation at UCSB. His research interests include nonlinear dynamical systems and control with application to aerospace and related systems. Dr. Teel has received the National Science Foundation Research Initiation and CAREER Awards, the 1998 IEEE Leon K. Kirchmayer Prize Paper Award, the 1998 George S. Axelby Outstanding Paper Award, and the first SIAM Control and Systems Theory Prize, in He was also the recipient of the 1999 Donald P. Eckman Award and the 2001 O. Hugo Schuck Best Paper Award, both given by the American Automatic Control Council.

Output Input Stability and Minimum-Phase Nonlinear Systems

Output Input Stability and Minimum-Phase Nonlinear Systems 422 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 47, NO. 3, MARCH 2002 Output Input Stability and Minimum-Phase Nonlinear Systems Daniel Liberzon, Member, IEEE, A. Stephen Morse, Fellow, IEEE, and Eduardo

More information

1 Lyapunov theory of stability

1 Lyapunov theory of stability M.Kawski, APM 581 Diff Equns Intro to Lyapunov theory. November 15, 29 1 1 Lyapunov theory of stability Introduction. Lyapunov s second (or direct) method provides tools for studying (asymptotic) stability

More information

On reduction of differential inclusions and Lyapunov stability

On reduction of differential inclusions and Lyapunov stability 1 On reduction of differential inclusions and Lyapunov stability Rushikesh Kamalapurkar, Warren E. Dixon, and Andrew R. Teel arxiv:1703.07071v5 [cs.sy] 25 Oct 2018 Abstract In this paper, locally Lipschitz

More information

THE area of robust feedback stabilization for general

THE area of robust feedback stabilization for general IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 52, NO. 11, NOVEMBER 2007 2103 Hybrid Feedback Control Robust Stabilization of Nonlinear Systems Christophe Prieur, Rafal Goebel, Andrew R. Teel Abstract In

More information

NOWADAYS, many control applications have some control

NOWADAYS, many control applications have some control 1650 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL 49, NO 10, OCTOBER 2004 Input Output Stability Properties of Networked Control Systems D Nešić, Senior Member, IEEE, A R Teel, Fellow, IEEE Abstract Results

More information

FOR linear time-invariant systems with outputs, there are

FOR linear time-invariant systems with outputs, there are 154 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 50, NO. 2, FEBRUARY 2005 Nonlinear Norm-Observability Notions Stability of Switched Systems João P. Hespanha, Senior Member, IEEE, Daniel Liberzon, Senior

More information

Stability theory is a fundamental topic in mathematics and engineering, that include every

Stability theory is a fundamental topic in mathematics and engineering, that include every Stability Theory Stability theory is a fundamental topic in mathematics and engineering, that include every branches of control theory. For a control system, the least requirement is that the system is

More information

1030 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 56, NO. 5, MAY 2011

1030 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 56, NO. 5, MAY 2011 1030 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL 56, NO 5, MAY 2011 L L 2 Low-Gain Feedback: Their Properties, Characterizations Applications in Constrained Control Bin Zhou, Member, IEEE, Zongli Lin,

More information

IN this paper we consider the stabilization problem for

IN this paper we consider the stabilization problem for 614 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL 42, NO 5, MAY 1997 Exponential Stabilization of Driftless Nonlinear Control Systems Using Homogeneous Feedback Robert T M Closkey, Member, IEEE, and Richard

More information

IN THIS paper we will consider nonlinear systems of the

IN THIS paper we will consider nonlinear systems of the IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 44, NO. 1, JANUARY 1999 3 Robust Stabilization of Nonlinear Systems Pointwise Norm-Bounded Uncertainties: A Control Lyapunov Function Approach Stefano Battilotti,

More information

Near-Potential Games: Geometry and Dynamics

Near-Potential Games: Geometry and Dynamics Near-Potential Games: Geometry and Dynamics Ozan Candogan, Asuman Ozdaglar and Pablo A. Parrilo January 29, 2012 Abstract Potential games are a special class of games for which many adaptive user dynamics

More information

Convergence Rate of Nonlinear Switched Systems

Convergence Rate of Nonlinear Switched Systems Convergence Rate of Nonlinear Switched Systems Philippe JOUAN and Saïd NACIRI arxiv:1511.01737v1 [math.oc] 5 Nov 2015 January 23, 2018 Abstract This paper is concerned with the convergence rate of the

More information

arxiv: v1 [math.oc] 30 May 2014

arxiv: v1 [math.oc] 30 May 2014 When is a Parameterized Controller Suitable for Adaptive Control? arxiv:1405.7921v1 [math.oc] 30 May 2014 Romeo Ortega and Elena Panteley Laboratoire des Signaux et Systèmes, CNRS SUPELEC, 91192 Gif sur

More information

Part III. 10 Topological Space Basics. Topological Spaces

Part III. 10 Topological Space Basics. Topological Spaces Part III 10 Topological Space Basics Topological Spaces Using the metric space results above as motivation we will axiomatize the notion of being an open set to more general settings. Definition 10.1.

More information

Approximation Metrics for Discrete and Continuous Systems

Approximation Metrics for Discrete and Continuous Systems University of Pennsylvania ScholarlyCommons Departmental Papers (CIS) Department of Computer & Information Science May 2007 Approximation Metrics for Discrete Continuous Systems Antoine Girard University

More information

Part V. 17 Introduction: What are measures and why measurable sets. Lebesgue Integration Theory

Part V. 17 Introduction: What are measures and why measurable sets. Lebesgue Integration Theory Part V 7 Introduction: What are measures and why measurable sets Lebesgue Integration Theory Definition 7. (Preliminary). A measure on a set is a function :2 [ ] such that. () = 2. If { } = is a finite

More information

A LaSalle version of Matrosov theorem

A LaSalle version of Matrosov theorem 5th IEEE Conference on Decision Control European Control Conference (CDC-ECC) Orlo, FL, USA, December -5, A LaSalle version of Matrosov theorem Alessro Astolfi Laurent Praly Abstract A weak version of

More information

ADAPTIVE control of uncertain time-varying plants is a

ADAPTIVE control of uncertain time-varying plants is a IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 56, NO. 1, JANUARY 2011 27 Supervisory Control of Uncertain Linear Time-Varying Systems Linh Vu, Member, IEEE, Daniel Liberzon, Senior Member, IEEE Abstract

More information

Lyapunov Stability Analysis of a Twisting Based Control Algorithm for Systems with Unmatched Perturbations

Lyapunov Stability Analysis of a Twisting Based Control Algorithm for Systems with Unmatched Perturbations 5th IEEE Conference on Decision and Control and European Control Conference (CDC-ECC) Orlando, FL, USA, December -5, Lyapunov Stability Analysis of a Twisting Based Control Algorithm for Systems with Unmatched

More information

FOR OVER 50 years, control engineers have appreciated

FOR OVER 50 years, control engineers have appreciated IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 49, NO. 7, JULY 2004 1081 Further Results on Robustness of (Possibly Discontinuous) Sample Hold Feedback Christopher M. Kellett, Member, IEEE, Hyungbo Shim,

More information

Distributed Coordinated Tracking With Reduced Interaction via a Variable Structure Approach Yongcan Cao, Member, IEEE, and Wei Ren, Member, IEEE

Distributed Coordinated Tracking With Reduced Interaction via a Variable Structure Approach Yongcan Cao, Member, IEEE, and Wei Ren, Member, IEEE IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 57, NO. 1, JANUARY 2012 33 Distributed Coordinated Tracking With Reduced Interaction via a Variable Structure Approach Yongcan Cao, Member, IEEE, and Wei Ren,

More information

IN THIS paper, we study the problem of asymptotic stabilization

IN THIS paper, we study the problem of asymptotic stabilization IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL 49, NO 11, NOVEMBER 2004 1975 Nonlinear Control of Feedforward Systems With Bounded Signals Georgia Kaliora and Alessandro Astolfi Abstract The stabilization

More information

Computational Tasks and Models

Computational Tasks and Models 1 Computational Tasks and Models Overview: We assume that the reader is familiar with computing devices but may associate the notion of computation with specific incarnations of it. Our first goal is to

More information

State and Parameter Estimation Based on Filtered Transformation for a Class of Second-Order Systems

State and Parameter Estimation Based on Filtered Transformation for a Class of Second-Order Systems State and Parameter Estimation Based on Filtered Transformation for a Class of Second-Order Systems Mehdi Tavan, Kamel Sabahi, and Saeid Hoseinzadeh Abstract This paper addresses the problem of state and

More information

Minimum-Phase Property of Nonlinear Systems in Terms of a Dissipation Inequality

Minimum-Phase Property of Nonlinear Systems in Terms of a Dissipation Inequality Minimum-Phase Property of Nonlinear Systems in Terms of a Dissipation Inequality Christian Ebenbauer Institute for Systems Theory in Engineering, University of Stuttgart, 70550 Stuttgart, Germany ce@ist.uni-stuttgart.de

More information

An asymptotic ratio characterization of input-to-state stability

An asymptotic ratio characterization of input-to-state stability 1 An asymptotic ratio characterization of input-to-state stability Daniel Liberzon and Hyungbo Shim Abstract For continuous-time nonlinear systems with inputs, we introduce the notion of an asymptotic

More information

RECENTLY, there has been renewed research interest

RECENTLY, there has been renewed research interest IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL 49, NO 12, DECEMBER 2004 2113 Distributed Control of Heterogeneous Systems Geir E Dullerud Raffaello D Andrea Abstract This paper considers control design for

More information

A Systematic Approach to Extremum Seeking Based on Parameter Estimation

A Systematic Approach to Extremum Seeking Based on Parameter Estimation 49th IEEE Conference on Decision and Control December 15-17, 21 Hilton Atlanta Hotel, Atlanta, GA, USA A Systematic Approach to Extremum Seeking Based on Parameter Estimation Dragan Nešić, Alireza Mohammadi

More information

LECTURE 10: REVIEW OF POWER SERIES. 1. Motivation

LECTURE 10: REVIEW OF POWER SERIES. 1. Motivation LECTURE 10: REVIEW OF POWER SERIES By definition, a power series centered at x 0 is a series of the form where a 0, a 1,... and x 0 are constants. For convenience, we shall mostly be concerned with the

More information

Linear robust output feedback control for permanent magnet synchronous motors with unknown load

Linear robust output feedback control for permanent magnet synchronous motors with unknown load Linear robust output feedback control for permanent magnet synchronous motors with unknown load Antonio Loria To cite this version: Antonio Loria. Linear robust output feedback control for permanent magnet

More information

AQUANTIZER is a device that converts a real-valued

AQUANTIZER is a device that converts a real-valued 830 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL 57, NO 4, APRIL 2012 Input to State Stabilizing Controller for Systems With Coarse Quantization Yoav Sharon, Member, IEEE, Daniel Liberzon, Senior Member,

More information

Observations on the Stability Properties of Cooperative Systems

Observations on the Stability Properties of Cooperative Systems 1 Observations on the Stability Properties of Cooperative Systems Oliver Mason and Mark Verwoerd Abstract We extend two fundamental properties of positive linear time-invariant (LTI) systems to homogeneous

More information

IN THIS PAPER, we consider a class of continuous-time recurrent

IN THIS PAPER, we consider a class of continuous-time recurrent IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II: EXPRESS BRIEFS, VOL. 51, NO. 4, APRIL 2004 161 Global Output Convergence of a Class of Continuous-Time Recurrent Neural Networks With Time-Varying Thresholds

More information

A framework for stabilization of nonlinear sampled-data systems based on their approximate discrete-time models

A framework for stabilization of nonlinear sampled-data systems based on their approximate discrete-time models 1 A framework for stabilization of nonlinear sampled-data systems based on their approximate discrete-time models D.Nešić and A.R.Teel Abstract A unified framework for design of stabilizing controllers

More information

Navigation and Obstacle Avoidance via Backstepping for Mechanical Systems with Drift in the Closed Loop

Navigation and Obstacle Avoidance via Backstepping for Mechanical Systems with Drift in the Closed Loop Navigation and Obstacle Avoidance via Backstepping for Mechanical Systems with Drift in the Closed Loop Jan Maximilian Montenbruck, Mathias Bürger, Frank Allgöwer Abstract We study backstepping controllers

More information

Memoryless output feedback nullification and canonical forms, for time varying systems

Memoryless output feedback nullification and canonical forms, for time varying systems Memoryless output feedback nullification and canonical forms, for time varying systems Gera Weiss May 19, 2005 Abstract We study the possibility of nullifying time-varying systems with memoryless output

More information

Stability Analysis and Synthesis for Scalar Linear Systems With a Quantized Feedback

Stability Analysis and Synthesis for Scalar Linear Systems With a Quantized Feedback IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL 48, NO 9, SEPTEMBER 2003 1569 Stability Analysis and Synthesis for Scalar Linear Systems With a Quantized Feedback Fabio Fagnani and Sandro Zampieri Abstract

More information

Characterization of Semantics for Argument Systems

Characterization of Semantics for Argument Systems Characterization of Semantics for Argument Systems Philippe Besnard and Sylvie Doutre IRIT Université Paul Sabatier 118, route de Narbonne 31062 Toulouse Cedex 4 France besnard, doutre}@irit.fr Abstract

More information

Near-Potential Games: Geometry and Dynamics

Near-Potential Games: Geometry and Dynamics Near-Potential Games: Geometry and Dynamics Ozan Candogan, Asuman Ozdaglar and Pablo A. Parrilo September 6, 2011 Abstract Potential games are a special class of games for which many adaptive user dynamics

More information

Math 350 Fall 2011 Notes about inner product spaces. In this notes we state and prove some important properties of inner product spaces.

Math 350 Fall 2011 Notes about inner product spaces. In this notes we state and prove some important properties of inner product spaces. Math 350 Fall 2011 Notes about inner product spaces In this notes we state and prove some important properties of inner product spaces. First, recall the dot product on R n : if x, y R n, say x = (x 1,...,

More information

Introduction to Real Analysis Alternative Chapter 1

Introduction to Real Analysis Alternative Chapter 1 Christopher Heil Introduction to Real Analysis Alternative Chapter 1 A Primer on Norms and Banach Spaces Last Updated: March 10, 2018 c 2018 by Christopher Heil Chapter 1 A Primer on Norms and Banach Spaces

More information

Passivity-based Stabilization of Non-Compact Sets

Passivity-based Stabilization of Non-Compact Sets Passivity-based Stabilization of Non-Compact Sets Mohamed I. El-Hawwary and Manfredi Maggiore Abstract We investigate the stabilization of closed sets for passive nonlinear systems which are contained

More information

CHAPTER 8: EXPLORING R

CHAPTER 8: EXPLORING R CHAPTER 8: EXPLORING R LECTURE NOTES FOR MATH 378 (CSUSM, SPRING 2009). WAYNE AITKEN In the previous chapter we discussed the need for a complete ordered field. The field Q is not complete, so we constructed

More information

1 Differentiable manifolds and smooth maps

1 Differentiable manifolds and smooth maps 1 Differentiable manifolds and smooth maps Last updated: April 14, 2011. 1.1 Examples and definitions Roughly, manifolds are sets where one can introduce coordinates. An n-dimensional manifold is a set

More information

Connectedness. Proposition 2.2. The following are equivalent for a topological space (X, T ).

Connectedness. Proposition 2.2. The following are equivalent for a topological space (X, T ). Connectedness 1 Motivation Connectedness is the sort of topological property that students love. Its definition is intuitive and easy to understand, and it is a powerful tool in proofs of well-known results.

More information

Monotone Control System. Brad C. Yu SEACS, National ICT Australia And RSISE, The Australian National University June, 2005

Monotone Control System. Brad C. Yu SEACS, National ICT Australia And RSISE, The Australian National University June, 2005 Brad C. Yu SEACS, National ICT Australia And RSISE, The Australian National University June, 005 Foreword The aim of this presentation is to give a (primitive) overview of monotone systems and monotone

More information

Abstract. Previous characterizations of iss-stability are shown to generalize without change to the

Abstract. Previous characterizations of iss-stability are shown to generalize without change to the On Characterizations of Input-to-State Stability with Respect to Compact Sets Eduardo D. Sontag and Yuan Wang Department of Mathematics, Rutgers University, New Brunswick, NJ 08903, USA Department of Mathematics,

More information

106 CHAPTER 3. TOPOLOGY OF THE REAL LINE. 2. The set of limit points of a set S is denoted L (S)

106 CHAPTER 3. TOPOLOGY OF THE REAL LINE. 2. The set of limit points of a set S is denoted L (S) 106 CHAPTER 3. TOPOLOGY OF THE REAL LINE 3.3 Limit Points 3.3.1 Main Definitions Intuitively speaking, a limit point of a set S in a space X is a point of X which can be approximated by points of S other

More information

Control of Mobile Robots

Control of Mobile Robots Control of Mobile Robots Regulation and trajectory tracking Prof. Luca Bascetta (luca.bascetta@polimi.it) Politecnico di Milano Dipartimento di Elettronica, Informazione e Bioingegneria Organization and

More information

ON THE REGULARITY OF SAMPLE PATHS OF SUB-ELLIPTIC DIFFUSIONS ON MANIFOLDS

ON THE REGULARITY OF SAMPLE PATHS OF SUB-ELLIPTIC DIFFUSIONS ON MANIFOLDS Bendikov, A. and Saloff-Coste, L. Osaka J. Math. 4 (5), 677 7 ON THE REGULARITY OF SAMPLE PATHS OF SUB-ELLIPTIC DIFFUSIONS ON MANIFOLDS ALEXANDER BENDIKOV and LAURENT SALOFF-COSTE (Received March 4, 4)

More information

AFAULT diagnosis procedure is typically divided into three

AFAULT diagnosis procedure is typically divided into three 576 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 47, NO. 4, APRIL 2002 A Robust Detection and Isolation Scheme for Abrupt and Incipient Faults in Nonlinear Systems Xiaodong Zhang, Marios M. Polycarpou,

More information

Filters in Analysis and Topology

Filters in Analysis and Topology Filters in Analysis and Topology David MacIver July 1, 2004 Abstract The study of filters is a very natural way to talk about convergence in an arbitrary topological space, and carries over nicely into

More information

Theorems. Theorem 1.11: Greatest-Lower-Bound Property. Theorem 1.20: The Archimedean property of. Theorem 1.21: -th Root of Real Numbers

Theorems. Theorem 1.11: Greatest-Lower-Bound Property. Theorem 1.20: The Archimedean property of. Theorem 1.21: -th Root of Real Numbers Page 1 Theorems Wednesday, May 9, 2018 12:53 AM Theorem 1.11: Greatest-Lower-Bound Property Suppose is an ordered set with the least-upper-bound property Suppose, and is bounded below be the set of lower

More information

Constrained Consensus and Optimization in Multi-Agent Networks

Constrained Consensus and Optimization in Multi-Agent Networks Constrained Consensus Optimization in Multi-Agent Networks The MIT Faculty has made this article openly available. Please share how this access benefits you. Your story matters. Citation As Published Publisher

More information

IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 57, NO. 11, NOVEMBER On the Performance of Sparse Recovery

IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 57, NO. 11, NOVEMBER On the Performance of Sparse Recovery IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 57, NO. 11, NOVEMBER 2011 7255 On the Performance of Sparse Recovery Via `p-minimization (0 p 1) Meng Wang, Student Member, IEEE, Weiyu Xu, and Ao Tang, Senior

More information

IN THIS paper we investigate the diagnosability of stochastic

IN THIS paper we investigate the diagnosability of stochastic 476 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL 50, NO 4, APRIL 2005 Diagnosability of Stochastic Discrete-Event Systems David Thorsley and Demosthenis Teneketzis, Fellow, IEEE Abstract We investigate

More information

1 Directional Derivatives and Differentiability

1 Directional Derivatives and Differentiability Wednesday, January 18, 2012 1 Directional Derivatives and Differentiability Let E R N, let f : E R and let x 0 E. Given a direction v R N, let L be the line through x 0 in the direction v, that is, L :=

More information

Nonlinear Discrete-Time Observer Design with Linearizable Error Dynamics

Nonlinear Discrete-Time Observer Design with Linearizable Error Dynamics 622 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 48, NO. 4, APRIL 2003 Nonlinear Discrete-Time Observer Design with Linearizable Error Dynamics MingQing Xiao, Nikolaos Kazantzis, Costas Kravaris, Arthur

More information

Propagating terraces and the dynamics of front-like solutions of reaction-diffusion equations on R

Propagating terraces and the dynamics of front-like solutions of reaction-diffusion equations on R Propagating terraces and the dynamics of front-like solutions of reaction-diffusion equations on R P. Poláčik School of Mathematics, University of Minnesota Minneapolis, MN 55455 Abstract We consider semilinear

More information

On finite gain L p stability of nonlinear sampled-data systems

On finite gain L p stability of nonlinear sampled-data systems Submitted for publication in Systems and Control Letters, November 6, 21 On finite gain L p stability of nonlinear sampled-data systems Luca Zaccarian Dipartimento di Informatica, Sistemi e Produzione

More information

2000 Mathematics Subject Classification. Primary: 37D25, 37C40. Abstract. This book provides a systematic introduction to smooth ergodic theory, inclu

2000 Mathematics Subject Classification. Primary: 37D25, 37C40. Abstract. This book provides a systematic introduction to smooth ergodic theory, inclu Lyapunov Exponents and Smooth Ergodic Theory Luis Barreira and Yakov B. Pesin 2000 Mathematics Subject Classification. Primary: 37D25, 37C40. Abstract. This book provides a systematic introduction to smooth

More information

The Dirichlet s P rinciple. In this lecture we discuss an alternative formulation of the Dirichlet problem for the Laplace equation:

The Dirichlet s P rinciple. In this lecture we discuss an alternative formulation of the Dirichlet problem for the Laplace equation: Oct. 1 The Dirichlet s P rinciple In this lecture we discuss an alternative formulation of the Dirichlet problem for the Laplace equation: 1. Dirichlet s Principle. u = in, u = g on. ( 1 ) If we multiply

More information

An introduction to Mathematical Theory of Control

An introduction to Mathematical Theory of Control An introduction to Mathematical Theory of Control Vasile Staicu University of Aveiro UNICA, May 2018 Vasile Staicu (University of Aveiro) An introduction to Mathematical Theory of Control UNICA, May 2018

More information

Distributed Control of Spatially Invariant Systems

Distributed Control of Spatially Invariant Systems IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 47, NO. 7, JULY 2002 1091 Distributed Control of Spatially Invariant Systems Bassam Bamieh, Member, IEEE, Fernando Paganini, Member, IEEE, and Munther A. Dahleh,

More information

Tree sets. Reinhard Diestel

Tree sets. Reinhard Diestel 1 Tree sets Reinhard Diestel Abstract We study an abstract notion of tree structure which generalizes treedecompositions of graphs and matroids. Unlike tree-decompositions, which are too closely linked

More information

Distributed Randomized Algorithms for the PageRank Computation Hideaki Ishii, Member, IEEE, and Roberto Tempo, Fellow, IEEE

Distributed Randomized Algorithms for the PageRank Computation Hideaki Ishii, Member, IEEE, and Roberto Tempo, Fellow, IEEE IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 55, NO. 9, SEPTEMBER 2010 1987 Distributed Randomized Algorithms for the PageRank Computation Hideaki Ishii, Member, IEEE, and Roberto Tempo, Fellow, IEEE Abstract

More information

Global stabilization of feedforward systems with exponentially unstable Jacobian linearization

Global stabilization of feedforward systems with exponentially unstable Jacobian linearization Global stabilization of feedforward systems with exponentially unstable Jacobian linearization F Grognard, R Sepulchre, G Bastin Center for Systems Engineering and Applied Mechanics Université catholique

More information

Foundations of Analysis. Joseph L. Taylor. University of Utah

Foundations of Analysis. Joseph L. Taylor. University of Utah Foundations of Analysis Joseph L. Taylor University of Utah Contents Preface vii Chapter 1. The Real Numbers 1 1.1. Sets and Functions 2 1.2. The Natural Numbers 8 1.3. Integers and Rational Numbers 16

More information

Nonlinear Tracking Control of Underactuated Surface Vessel

Nonlinear Tracking Control of Underactuated Surface Vessel American Control Conference June -. Portland OR USA FrB. Nonlinear Tracking Control of Underactuated Surface Vessel Wenjie Dong and Yi Guo Abstract We consider in this paper the tracking control problem

More information

Chapter III. Stability of Linear Systems

Chapter III. Stability of Linear Systems 1 Chapter III Stability of Linear Systems 1. Stability and state transition matrix 2. Time-varying (non-autonomous) systems 3. Time-invariant systems 1 STABILITY AND STATE TRANSITION MATRIX 2 In this chapter,

More information

On integral-input-to-state stabilization

On integral-input-to-state stabilization On integral-input-to-state stabilization Daniel Liberzon Dept. of Electrical Eng. Yale University New Haven, CT 652 liberzon@@sysc.eng.yale.edu Yuan Wang Dept. of Mathematics Florida Atlantic University

More information

Metric Spaces and Topology

Metric Spaces and Topology Chapter 2 Metric Spaces and Topology From an engineering perspective, the most important way to construct a topology on a set is to define the topology in terms of a metric on the set. This approach underlies

More information

Hybrid Systems Techniques for Convergence of Solutions to Switching Systems

Hybrid Systems Techniques for Convergence of Solutions to Switching Systems Hybrid Systems Techniques for Convergence of Solutions to Switching Systems Rafal Goebel, Ricardo G. Sanfelice, and Andrew R. Teel Abstract Invariance principles for hybrid systems are used to derive invariance

More information

Global Analysis of Piecewise Linear Systems Using Impact Maps and Surface Lyapunov Functions

Global Analysis of Piecewise Linear Systems Using Impact Maps and Surface Lyapunov Functions IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL 48, NO 12, DECEMBER 2003 2089 Global Analysis of Piecewise Linear Systems Using Impact Maps and Surface Lyapunov Functions Jorge M Gonçalves, Alexandre Megretski,

More information

6 Lecture 6: More constructions with Huber rings

6 Lecture 6: More constructions with Huber rings 6 Lecture 6: More constructions with Huber rings 6.1 Introduction Recall from Definition 5.2.4 that a Huber ring is a commutative topological ring A equipped with an open subring A 0, such that the subspace

More information

Global Attractors in PDE

Global Attractors in PDE CHAPTER 14 Global Attractors in PDE A.V. Babin Department of Mathematics, University of California, Irvine, CA 92697-3875, USA E-mail: ababine@math.uci.edu Contents 0. Introduction.............. 985 1.

More information

Finite-Time Behavior of Inner Systems

Finite-Time Behavior of Inner Systems 1134 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 48, NO. 7, JULY 2003 Finite-Time Behavior of Inner Systems Jobert H. A. Ludlage, Member, IEEE, Siep Weiland, Anton A. Stoorvogel, Senior Member, IEEE,

More information

ECE7850 Lecture 8. Nonlinear Model Predictive Control: Theoretical Aspects

ECE7850 Lecture 8. Nonlinear Model Predictive Control: Theoretical Aspects ECE7850 Lecture 8 Nonlinear Model Predictive Control: Theoretical Aspects Model Predictive control (MPC) is a powerful control design method for constrained dynamical systems. The basic principles and

More information

Chapter One. Introduction

Chapter One. Introduction Chapter One Introduction A system is a combination of components or parts that is perceived as a single entity. The parts making up the system may be clearly or vaguely defined. These parts are related

More information

Strong Lyapunov Functions for Systems Satisfying the Conditions of La Salle

Strong Lyapunov Functions for Systems Satisfying the Conditions of La Salle 06 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 49, NO. 6, JUNE 004 Strong Lyapunov Functions or Systems Satisying the Conditions o La Salle Frédéric Mazenc and Dragan Ne sić Abstract We present a construction

More information

IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 50, NO. 5, MAY Bo Yang, Student Member, IEEE, and Wei Lin, Senior Member, IEEE (1.

IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 50, NO. 5, MAY Bo Yang, Student Member, IEEE, and Wei Lin, Senior Member, IEEE (1. IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL 50, NO 5, MAY 2005 619 Robust Output Feedback Stabilization of Uncertain Nonlinear Systems With Uncontrollable and Unobservable Linearization Bo Yang, Student

More information

Robust Stabilization of Non-Minimum Phase Nonlinear Systems Using Extended High Gain Observers

Robust Stabilization of Non-Minimum Phase Nonlinear Systems Using Extended High Gain Observers 28 American Control Conference Westin Seattle Hotel, Seattle, Washington, USA June 11-13, 28 WeC15.1 Robust Stabilization of Non-Minimum Phase Nonlinear Systems Using Extended High Gain Observers Shahid

More information

Existence and Uniqueness

Existence and Uniqueness Chapter 3 Existence and Uniqueness An intellect which at a certain moment would know all forces that set nature in motion, and all positions of all items of which nature is composed, if this intellect

More information

We are going to discuss what it means for a sequence to converge in three stages: First, we define what it means for a sequence to converge to zero

We are going to discuss what it means for a sequence to converge in three stages: First, we define what it means for a sequence to converge to zero Chapter Limits of Sequences Calculus Student: lim s n = 0 means the s n are getting closer and closer to zero but never gets there. Instructor: ARGHHHHH! Exercise. Think of a better response for the instructor.

More information

CS264: Beyond Worst-Case Analysis Lecture #11: LP Decoding

CS264: Beyond Worst-Case Analysis Lecture #11: LP Decoding CS264: Beyond Worst-Case Analysis Lecture #11: LP Decoding Tim Roughgarden October 29, 2014 1 Preamble This lecture covers our final subtopic within the exact and approximate recovery part of the course.

More information

L 2 -induced Gains of Switched Systems and Classes of Switching Signals

L 2 -induced Gains of Switched Systems and Classes of Switching Signals L 2 -induced Gains of Switched Systems and Classes of Switching Signals Kenji Hirata and João P. Hespanha Abstract This paper addresses the L 2-induced gain analysis for switched linear systems. We exploit

More information

Richard DiSalvo. Dr. Elmer. Mathematical Foundations of Economics. Fall/Spring,

Richard DiSalvo. Dr. Elmer. Mathematical Foundations of Economics. Fall/Spring, The Finite Dimensional Normed Linear Space Theorem Richard DiSalvo Dr. Elmer Mathematical Foundations of Economics Fall/Spring, 20-202 The claim that follows, which I have called the nite-dimensional normed

More information

Immersion and Invariance: A New Tool for Stabilization and Adaptive Control of Nonlinear Systems

Immersion and Invariance: A New Tool for Stabilization and Adaptive Control of Nonlinear Systems 590 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 48, NO. 4, APRIL 2003 Immersion and Invariance: A New Tool for Stabilization and Adaptive Control of Nonlinear Systems Alessandro Astolfi, Senior Member,

More information

Decomposing Bent Functions

Decomposing Bent Functions 2004 IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 49, NO. 8, AUGUST 2003 Decomposing Bent Functions Anne Canteaut and Pascale Charpin Abstract In a recent paper [1], it is shown that the restrictions

More information

642:550, Summer 2004, Supplement 6 The Perron-Frobenius Theorem. Summer 2004

642:550, Summer 2004, Supplement 6 The Perron-Frobenius Theorem. Summer 2004 642:550, Summer 2004, Supplement 6 The Perron-Frobenius Theorem. Summer 2004 Introduction Square matrices whose entries are all nonnegative have special properties. This was mentioned briefly in Section

More information

Standard forms for writing numbers

Standard forms for writing numbers Standard forms for writing numbers In order to relate the abstract mathematical descriptions of familiar number systems to the everyday descriptions of numbers by decimal expansions and similar means,

More information

A Generalized Uncertainty Principle and Sparse Representation in Pairs of Bases

A Generalized Uncertainty Principle and Sparse Representation in Pairs of Bases 2558 IEEE TRANSACTIONS ON INFORMATION THEORY, VOL 48, NO 9, SEPTEMBER 2002 A Generalized Uncertainty Principle Sparse Representation in Pairs of Bases Michael Elad Alfred M Bruckstein Abstract An elementary

More information

Infinite-dimensional perturbations, maximally nondensely defined symmetric operators, and some matrix representations

Infinite-dimensional perturbations, maximally nondensely defined symmetric operators, and some matrix representations Available online at www.sciencedirect.com ScienceDirect Indagationes Mathematicae 23 (2012) 1087 1117 www.elsevier.com/locate/indag Infinite-dimensional perturbations, maximally nondensely defined symmetric

More information

SPARSE signal representations have gained popularity in recent

SPARSE signal representations have gained popularity in recent 6958 IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 57, NO. 10, OCTOBER 2011 Blind Compressed Sensing Sivan Gleichman and Yonina C. Eldar, Senior Member, IEEE Abstract The fundamental principle underlying

More information

Hausdorff Measure. Jimmy Briggs and Tim Tyree. December 3, 2016

Hausdorff Measure. Jimmy Briggs and Tim Tyree. December 3, 2016 Hausdorff Measure Jimmy Briggs and Tim Tyree December 3, 2016 1 1 Introduction In this report, we explore the the measurement of arbitrary subsets of the metric space (X, ρ), a topological space X along

More information

Solution of the 8 th Homework

Solution of the 8 th Homework Solution of the 8 th Homework Sangchul Lee December 8, 2014 1 Preinary 1.1 A simple remark on continuity The following is a very simple and trivial observation. But still this saves a lot of words in actual

More information

IN this paper, we consider the capacity of sticky channels, a

IN this paper, we consider the capacity of sticky channels, a 72 IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 54, NO. 1, JANUARY 2008 Capacity Bounds for Sticky Channels Michael Mitzenmacher, Member, IEEE Abstract The capacity of sticky channels, a subclass of insertion

More information

ASIGNIFICANT research effort has been devoted to the. Optimal State Estimation for Stochastic Systems: An Information Theoretic Approach

ASIGNIFICANT research effort has been devoted to the. Optimal State Estimation for Stochastic Systems: An Information Theoretic Approach IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL 42, NO 6, JUNE 1997 771 Optimal State Estimation for Stochastic Systems: An Information Theoretic Approach Xiangbo Feng, Kenneth A Loparo, Senior Member, IEEE,

More information

Ordinary Differential Equation Introduction and Preliminaries

Ordinary Differential Equation Introduction and Preliminaries Ordinary Differential Equation Introduction and Preliminaries There are many branches of science and engineering where differential equations arise naturally. Now days, it finds applications in many areas

More information

Converse Lyapunov theorem and Input-to-State Stability

Converse Lyapunov theorem and Input-to-State Stability Converse Lyapunov theorem and Input-to-State Stability April 6, 2014 1 Converse Lyapunov theorem In the previous lecture, we have discussed few examples of nonlinear control systems and stability concepts

More information