![MathJax Logo](/templates/jsp/_style2/_tandf/pb2/images/math-jax.gif)
ABSTRACT
When observabilities of hybrid dynamic systems are considered, the distinguishability of subsystems takes a very important role. Necessary and sufficient conditions for distinguishability of linear dynamic systems are obtained. Some illustrative examples are presented.
1. Introduction
The switched system is an important case of a hybrid system. As a special kind of linear switched systems has been extensively investigated [Citation1–5]. When we consider the observability of switched systems composed by time-invariant subsystems, distinguishability plays a crucial role (see [Citation6]). Among the references about distinguishability of hybrid systems, we would like to refer the readers to the papers [Citation7–9]. Distinguishability of switched systems is concerned with recovering the initial state as well as the switching signal from the output (and input) and has been widely studied, see e.g. [Citation10] for continuous linear control homogenous systems, Lou et al. [Citation4] for continuous linear control inhomogeneous systems and recently, relatively easy equivalent conditions to verify distinguishability are presented in [Citation3]. For discrete switched case, Baglietto et al. [Citation1,Citation2], considered the problem of identifying a discrete-time nonlinear system, within a finite family of possible models, from data sequences of a finite length. The problem is approached by resorting to the notion of output distinguishability.
However, there are applications of switched systems where their temporal nature cannot be represented by the continuous line or a discrete uniform time domain [Citation11,Citation12]. Indeed, a closed-loop system consisting of a continuous-time system and an intermittent controller is one application [Citation13,Citation14]. The consensus problem under intermittent information due to communication obstacles and limitations of sensors is another example.
Time scale theory is very useful since it is an appropriate tool to study continuous and discrete-time systems in a uniform framework [Citation7,Citation15–18]. The objective of this paper is to extend the distinguishability results of a class of continuous-time linear control switched systems in [Citation3,Citation4]. Necessary and sufficient conditions for input distinguishability of linear switched dynamic systems on time scales.
The rest of the paper is organized as follows. Section 2 recalls some preliminaries on time scale theory. The studied class of systems, input distinguishability concept, necessary and sufficient conditions for input distinguishability for linear control dynamic switched systems are obtained in Section 3.
2 Preliminaries
We recall some basics on time scale theory (for more details see [Citation16,Citation17]). A nonempty subset of real line is called time scales and it is denoted by
.
Let be a time scale. As usual, for
,
,
,
and
define and denote its forward jump operator, backward jump operator, forward graininess function and backward graininess function.
A point , is called right-scattered (right-dense), left-scattered (left-dense) and isolated (dense), if
,
and
, respectively.
A set is derived from a time scale
: if
has a left-scattered maximum
, then
, otherwise
.
Given a time scale interval , then
denotes the interval
if
and denotes the interval
if
. In fact,
. Also, for
, we define
. If
is a bounded time scale, then
can be identified with
.
If and
, then we define the following neighbourhoods of
,
and
.
Let us consider some examples of time scales (see [Citation17]).
If
,
is a time scale. Then we have
and
for all
. Hence each point
is isolated,
for all
and
.
Let
. Then
A function is called rd-continuous provided that it is continuous at right-dense points in
, and has a finite limit at left-dense points, and the set of rd-continuous functions are denoted by
. The set of functions
includes the functions
whose derivative is in
too.
It is known [Citation9] that for every there exists at least one partition
of
such that for each
either
or
and
. For given
we denote by
the set of all partitions
that possess the above property.
Let be a bounded function on
, and let
be a partition of
. In each interval
, where
, choose an arbitrary point
and form the sum
We call
a Riemann
-sum of
corresponding to the partition
.
We say that is Riemann
-integrable from
to
(or on
) if there exists a number
with the following property: for each
there exists
such that
for every Riemann
-sum
of
corresponding to a partition
independent of the way in which we choose
. It is easily seen that such a number
is unique. The number
is the Riemann
-integral of
from
to
, and we will denote it by
.
A bounded function is Riemann
-integrable on
if and only if the set of all right-dense points of
at which
is discontinuous is a set of
-measure zero.
Assume that ,
and
is rd-continuous. Then the integral has the following properties.
If
, then
, where the integral on the right-hand side is the Riemann integral.
If
consists of isolated points, then
Let be Riemann
-integrable function on
. If
has a
-antiderivative
, then
. In particular,
for all
(see [Citation17, Theorem 1.75]).
Let be a function which is Riemann
-integrable from
to
. For
, let
. Then
is continuous on
. Further, let
and let
be arbitrary at
if
is right-scattered, and let
be continuous at
if
is right-dense. Then
is
-differentiable at
and
(see ([Citation8, Theorem 4.3])).
A function is called regressive if
is invertible for all
. The set of regressive functions is denoted by
(or shortly denoted by
). For simplicity, we denote by
the set of complex regressive constants and similarly, we define the set of
.
For definition of the exponential function on time scales see [Citation16]. A function is called to be positively regressive if
for all
. If
is a positively regressive function and
, then (see [Citation16]) the exponential function
is the unique solution of the initial value problem
In particular, if
is such that
for all
, we have
if
,
if
with
.
The definition of generalized monomials on time scales (see [Citation17, Section 1.6]) is given as
for
. It follows that
where
denotes
-derivative of the
with respect to
.
Using induction, it is easy to see that holds for all
and all
with
and
holds for all
and all
with
.
Throughout this paper, we assume that . The next definitions and results were given in [Citation19].
Let . A function
has exponential order
on
, if
and there exists
such that
for all
. For
, it is easy to see that
is of exponential order
on
.
Let for
. If
with
for all
, then
for all
, so that
is of exponential order
.
Now we give the definition of the Laplace transform (see [Citation19, Definition 4.5]):
Let be a function. Then the Laplace transform
about the point
of the function
is defined by
(1)
(1) where
such that
for which the improper integral converges.
For , let
and
. The minimal graininess function
by
and for
, define
By using equation (1),
The following results about the Laplace transform were proved in [Citation19].
Theorem 2.1:
Let be of exponential order
. Then the Laplace transform
exists on
and converges absolutely.
Theorem 2.2:
Let be of exponential order
. Then the Laplace transform
converges uniformly in the half-plane
, where
.
Theorem 2.3:
Let be of exponential order
, respectively. Then for any
, we have
on
.
Recently, Zada et al. [Citation20], proved the following spectral decomposition theorem on time scales:
Let be a regressive matrix of order
. For each
there exist
such that
Moreover, if
then
for all
and there exist
-valued polynomials
with
such that
It follows that
(2)
(2) By using Theorem 2.3 and first translation theorem (see [Citation21, Theorem 3.2]) of the Laplace transform, the form (2) becomes proper rational function.
On the other hand, if we have a proper rational function, then by using partial fractions method and applying Laplace inverse it is easy to get the form (2).
For various properties of the Laplace transform on time scales, we refer to [Citation16,Citation17,Citation19,Citation21].
Before addressing the problem of distinguishability, it helps to recall the solution frame work assumptions with the following time-invariant dynamic linear system
(3)
(3) where
are the state and input vectors and
,
are constant matrices.
For a regressive matrix and rd-continuous input
, the time-invariant dynamic linear system (3) with initial condition
has a unique solution of the form
3 Distinguishability
Consider a switched dynamic system composed by time-invariant subsystems ():
(4)
(4) where
,
and
. Naturally,
(5)
(5) Without loss of generality, we can assume only two subsystems i.e.
. Denote
(6)
(6) and
(7)
(7) Recently, in [Citation4], the authors gave a notion of distinguishability for linear non-autonomous systems and yielded a necessary and sufficient condition for distinguishability of two linear systems.
We refer the reader to [Citation8,Citation9,Citation16,Citation22] for a broad introduction to -measure and integration theory.
For -measurable set
, a
-measurable vector-valued function
belongs to
provided that
Definition 3.1:
(see [Citation4]) Let and
. We say that
and
are said to be distinguishable on
if for any non-zero
the corresponding outputs
and
cannot be identical to each other on
.
To study the distinguishability of two subsystems, some auxiliary concepts of distinguishability have been stated here:
Definition 3.2:
(see [Citation4]) Let be a function space. We say that
and
are
input distinguishable on
if for any non-zero
the outputs
and
cannot be identical to each other on
.
Especially, when is the set of generalized polynomial function class, the set of analytic function class and the set of smooth function class
, then the corresponding distinguishability is called “generalized polynomial input distinguishability”, “analytic input distinguishability” and “smooth input distinguishability”, respectively.
The distinguishability of and
on
is equivalent to that for the following system:
(8)
(8)
implies that
on
Thus the problem relates to the notion of zero dynamics.
The closed form solution with initial condition of the system (8) is as follows
(9)
(9) Let us define the following infinite order matrices
(10)
(10) and
(11)
(11) Our first result characterized generalized polynomial input distinguishability:
Theorem 3.3:
The generalized polynomial input distinguishability of and
is independent of
which is equivalent to that of every sub-matrix composed of the left finite column vector of
has full column rank.
Moreover, the necessary and sufficient conditions for the -th generalized polynomial input distinguishability of
and
is that
has full column rank. While the necessary and sufficient conditions for the generalized polynomial input distinguishability of
and
is that for any
,
has full column rank.
Proof:
For the given input , the corresponding output of the system (8) is as follows
(12)
(12) Let
be an
-th
-valued generalized polynomial on
:
where
.
Therefore, we have
It follows that
is analytic. Therefore
holds if and only if all
-derivatives of
at
equal to zero:
(13)
(13) Then the equation (13), implies the following system of equations:
(14)
(14) We get that (14) is equivalent to the following infinite dimensional equation:
(15)
(15) Therefore, the
th generalized polynomial input distinguishability of
and
is equivalent to that (15) admits only trivial solution. That is to say
has full column rank.
Hence it is generalized polynomial input distinguishable and also independent of .
The following result is an immediate consequence of the above theorem and proof is similar to [Citation4 Corollary 3.2].
Corollary 3.4:
If and
are generalized polynomial input distinguishable, then
.
Generalizing the idea of Theorem 3.3, we can obtain the following result:
Theorem 3.5:
The analytic input distinguishability of and
on
are equivalent to the following infinite dimensional equation
(16)
(16) having only trivial solution such that the corresponding series
converges in an open interval including
.
To prove our next results, we need to recall some concepts from [Citation4].
It is well known that a special kind of -type matrix
is called
-type matrix if
It is easy to see that the product of two
-type matrices is still a
-type matrix.
In [Citation4], Lou et al. introduced three types of invertible transformations on a
-type matrix
:
Type I: Left-multiply by an invertible
-type matrix
.
Type II: If for some and
, all
th row vectors of
are zero, but the
th row vector is not zero, then replace
th row by
th row
.
Type III: If for some , all
th row vectors
are zero, then delete these rows.
Let us recall Lemma 4.6 of [Citation4]:
Lemma 3.6:
Let
be a
-type matrix. Then there exist an
, an invertible
matrix
and an invertible transform
, which is composed by finite transformations of types I-III, such that
(17)
(17) when
or
(18)
(18) when
, where
are
matrices
.
The following consequence of Lemma 3.6 is as follows. The proof is similar to the one in [Citation4, Theorem 4.5] and therefore omitted.
Theorem 3.7:
Let , then
and
are not analytic input distinguishable.
The following Lemma is corollary of Lemma 3.6:
Lemma 3.8:
Let . Suppose that
Then if the infinite order linear equation
admits non-trivial solutions, it must admit some non-trivial solution such that
converges in
.
Under the light of Lemma 3.8, we have the following interesting and important result.
Theorem 3.9:
The analytic input distinguishability of and
on
is equivalent to that (16) admits only trivial solution. Consequently, it is independent of
.
We will now show that Theorem 3.9 implies that the smooth input distinguishability and analytic input distinguishability are equivalent.
Theorem 3.10:
The analytic input distinguishability of and
is equivalent to the smooth input distinguishability of
and
.
Theorem 3.9 gives us conditions not only necessary but also sufficient to the smooth input distinguishability of the linear systems. However, the conditions are not easy enough to verify. Let us go further for an equivalent conditions which is relatively more easy to verify.
From Theorem 3.7, if and
are not analytic input distinguishable, then there exists a pair
such that
(19)
(19)
and
(20)
(20) with
(21)
(21) for some
. It follows that
and Laplace transform
can be defined for any
.
Denote
and consider the Laplace transform of
in (19), we obtain the following results. These results are a straightforward generalization from the linear ODE-case [Citation3, Lemma 3.1 and Lemma 3.2]:
Lemma 3.11:
If and
are not distinguishable, then we can find a pair
satisfying (19) with
where
and
are vector-valued polynomials
.
Lemma 3.12:
If and
are not distinguishable, then we can find a pair
satisfying (19) with
where
and
.
By using the above results, it implies that the necessary and sufficient conditions for -th generalized polynomial input distinguishability and the
-th generalized polynomial input distinguishable are equivalent. Hereafter, we can see that for any
, the matrix
has full column rank if and only if
(22)
(22) has full column rank.
From the classical Cayley–Hamilton theorem (22) is equivalent to that
has a full column rank.
For , consider
Similar to previous notations (see e.g. (6) and (7)), it follows that
(23)
(23) We claim for any
and
,
, the solution of (23) corresponding to
and
satisfies
on
. Equivalently
and
are
th polynomial input distinguishable.
If it is not the case, then we have such that the corresponding
. Let
Then
solves (8) with
Since
, by considering the real part or imaginary part of
and
, it follows that
and
are not distinguishable. This is a contradiction.
Therefore, Theorem 3.3, implies that
has full column rank.
Summarizing the above arguments, we obtain our main result.
Theorem 3.13:
Subsystems and
are analytic input distinguishable if and only if for any
, the matrix
has a full column rank.
Example 3.14:
Consider ,
with their matrices being
It is easy to see that
Then and
are generalized polynomial input distinguishable. However,
does not satisfy the full column rank condition for
. Therefore,
and
are not analytic input distinguishable. It implies that generalized polynomial input distinguishability is weaker than the analytic input distinguishability.
Acknowledgements
The authors are grateful to the referees for their careful reading of the manuscript and their valuable comments. We also thanks the editor.
Disclosure statement
No potential conflict of interest was reported by the authors.
ORCID
Awais Younus http://orcid.org/0000-0002-0590-6691
Muhammad Asif http://orcid.org/0000-0001-5229-5371
References
- Baglietto M, Battistelli G, Tesi P. Mode-observability degree in discrete-time switching linear systems. Syst Control Lett. 2014;70:69–76. doi: 10.1016/j.sysconle.2014.05.006
- Baglietto M, Battistelli G, Tesi P. Distinguishability of discrete-time nonlinear systems. IEEE Trans Autom Control. 2014;59(4):1014–1020. doi: 10.1109/TAC.2013.2283132
- Lou H. Necessary and sufficient conditions for distinguishability of linear control systems. Acta Math Appl Sin. 2014;30(2):473–482. doi: 10.1007/s10255-014-0283-1
- Lou H, Si P. The distinguishability of linear control systems. Nonlinear Anal: Hybrid Syst. 2009;3:21–38.
- Zada A, Zada B. On uniform exponential stability of linear switching system. Math Methods Appl Sci. 2019;42(2):717–722. doi: 10.1002/mma.5373
- Collins P, Van Schuppen JH. Observability of piecewise-affine hybrid systems. In: Hybrid systems: computation and control, lecture notes in computer science, 2993 (2004) 265–279.
- Eisenbarth G, Davis JM, Gravagne IA. Singular value conditions for stability of dynamic switched systems. J Math Anal Appl 2017;452(2):814–829. doi: 10.1016/j.jmaa.2017.02.059
- Guseinov GS. Integration on time scales. J Math Anal Appl. 2003;285:107–127. doi: 10.1016/S0022-247X(03)00361-5
- Guseinov GS, Kaymakcalan B. Basics of Riemann delta and nabla integration on time scales. J Differ Equ Appl. 2002;8:1001–1017. doi: 10.1080/10236190290015272
- Vidal R, Chiuso A, Soatto S, et al. Observability of linear hybrid systems. In: Hybrid systems: computation and control, lecture notes in computer science, 2623 (2003), 526–539.
- Taousser FZ, Defoort M, Djemai M, et al. Stability analysis of a class of switched nonlinear systems using the time scale theory. Nonlinear Anal: Hybrid Syst. 2019;33:195–210.
- Taousser FZ, Defoort M, Djemai M. Stability analysis of a class of switched linear systems on non-uniform time domains. Syst Control Lett. 2014;74:24–31. doi: 10.1016/j.sysconle.2014.09.012
- Tunç C, Tunç O. A note on the qualitative analysis of Volterra integro-differential equations. J Taibah Univ Sci. 2019;13(1):490–496. doi: 10.1080/16583655.2019.1596629
- Tunç C, Tunç O. On behaviours of functional Volterra integro-differential equations with multiple time lags. J Taibah Univ Sci. 2018;12(2):173–179. doi: 10.1080/16583655.2018.1451117
- Bejarano FJ. Reconstructability of controlled switched linear systems: discrete and continuous states. Int J Dyn Control. 2018;6(3):1218–1230. doi: 10.1007/s40435-016-0284-4
- Bohner M, Peterson A. Advanced in dynamic equations on time scale. Boston: Birkhauser; 2003.
- Bohner M, Peterson A. Dynamic equations on time scale, an introduction with applications. Boston: Birkhauser; 2001.
- Davis JM, Gravange IA, Marks RJ et al. Stability of switched linear systems on non-uniform time domains. In Proceedings of the 42nd South Eastern Symposium on System Theory, Tyler, TX, USA, March 7–8; 2010. p. 127–132.
- Bohner M, Guseinov GS, Karpuz B. Properties of the Laplace transform on time scales with arbitrary graininess. Integral Transforms Spec Funct. 2011;22(11):785–800. doi: 10.1080/10652469.2010.548335
- Zada A, Li T, Ismail S, et al. Exponential dichotomy of linear autonomous systems over the time scales. Differ Equ Appl. 2016;8(2):123–134.
- Bohner M, Guseinov GS, Karpuz B. Further properties of the Laplace transform on time scales with arbitrary graininess. Integral Transforms Spec Funct. 2013;24(4):289–301. doi: 10.1080/10652469.2012.689300
- Agarwal RP, Otero-Espinar V, Perera K, et al. Basic properties of Sobolev’s spaces on time scales. Adv Differ Equ. 2006; Art. ID 38121, 14 pp.