Hindawi Publishing Corporation

Mathematical Problems in Engineering

Volume 2015, Article ID 806706, 14 pages
http://dx.doi.org/10.1155/2015/806706

Research Article
Impulsive Controllability/Observability for Interconnected
Descriptor Systems with Two Subsystems

Qingling Zhang,1,2 Jian Li,1 and Junchao Ren1

1
Institute of Systems Science, Northeastern University, Shenyang, Liaoning 110819, China
2
State Key Laboratory of Synthetical Automation for Process Industries, Northeastern University, Shenyang, Liaoning 110819, China

Correspondence should be addressed to Qingling Zhang; qlzhang@mail.neu.edu.cn

Received 30 July 2015; Revised 12 November 2015; Accepted 3 December 2015

Academic Editor: Jean J. Loiseau

Copyright © 2015 Qingling Zhang et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The problem of decentralized impulse controllability/observability for large-scale interconnected descriptor systems with two
subsystems by derivative feedback is studied. Necessary conditions for the existence of a derivative feedback controller for the
first subsystem of the large-scale interconnected descriptor systems ensuring the second subsystem to be impulse controllable
and impulse observable are derived, respectively. Based on the results, a derivative feedback controller for the first subsystem of
the large-scale interconnected descriptor systems is constructed easily such that the second subsystem is impulse controllable or
impulse observable. Finally, examples are given to illustrate the effectiveness of the results obtained in this paper.

1. Introduction feedback transformations on descriptor linear systems, and

Descriptor systems are also referred to as singular systems,
implicit systems, differential algebraic systems, semistate sys-
tems, or generalized state space systems. Descriptor system
models are more convenient and natural than standard state
space system models in order to describe practical systems,
such as interconnected large-scale systems, economic sys-
tems, networks, power systems, and chemical processes [1, 2].
Due to this reason, the control of descriptor systems has been
extensively studied in past years and a great number of results
based on the theory of standard state space systems have been
generalized to descriptor systems [1–5].
It has been known that the response of a descriptor
system may contain impulsive terms. In a practical system,
the impulse may cause bad performance and even destroys
the system. Therefore, it is important to study the problem
of eliminating impulsive behavior of a descriptor systems via
certain feedback controllers. For regular descriptor systems, it
is pointed out in the work [6] that the necessary and sufficient
condition for impulse elimination via proportional state
feedback is that the considered system is impulse controllable.
The work in [7] shows that homogeneous indices are a
complete set of invariants for the action of a natural group of
the homogeneous indices determine exactly which closed-
loop invariant polynomials can be assigned by feedback,
thereby generalizing the control structure theorem of Rosen-
brock. In [8], the pole placement and connected problems by
the use of matrix pencil formulation have also been studied.
In [9], the limits in altering the eigenstructure of linear
reachable descriptor systems by proportional plus derivative
(P-D) state feedback have also been studied, and an explicit
necessary and sufficient condition is established for a set
of invariant polynomials and positive integers to represent
the finite and the infinite eigenstructure of a system obtain-
able from the given descriptor system by proportional plus
derivative (P-D) state feedback. The problem of pole structure
assignment (PSA) by proportional state feedback in implicit,
linear, and uncontrollable systems has also been studied in
[10]. In [11], the problem of the action of proportional plus
derivative (P-D) and pure proportional feedback groups on
the set of all singular systems has also been studied. In
[12], the concepts of impulsive modes and impulsive mode
controllability are proposed by the analysis of admissibility
of initial conditions. In [13], the problem of impulsive mode
elimination via proportional state feedback is considered.
It is pointed out that the impulsive mode controllability
2 Mathematical Problems in Engineering
is necessary and sufficient for the existence of impulsive
mode eliminating proportional state feedback controller. The
concept of impulsive controllability is proposed in [14]. It is
shown [15, 16] that the condition of impulse elimination via
proportional plus derivative (P-D) state feedback is that the
original system is impulsive controllable one. New criteria
are proposed in [17] for impulsive mode controllability
of descriptor linear systems by adopting the null space
approach. In [18], the problem of impulse elimination via pro-
portional plus derivative (P-D) output feedback has also been
studied. In [19], a decentralized (P-D) feedback is explored for
the impulsive elimination problem where the centralized (P-
D) feedback is regarded as a special case. In [20], a concept
of the structured proportional plus derivative (P-D) feedback
is first introduced and an explicit necessary and sufficient
condition is constructively derived for the closed-loop system
to be regular and impulse-free by the structured proportional
plus derivative (P-D) feedback. A rank condition of impulse
observability is presented for regular descriptor systems [21].
In [12], the presence of impulsive responses in descriptor
systems and how they relate to impulse controllability and
impulse observability are considered.
The large-scale interconnected descriptor systems are
a kind of dynamic systems which are more general and
have extensive applied background, such as power systems,
economic systems, and network systems [22]. The class of
systems can be characterized by a large number of state
variables, parametric uncertainties, a complex structure, and
a strong interaction between subsystems [23]. Decentral-
ized/interconnected descriptor systems have attracted more
and more attention because of their practical backgrounds;
for example, [24, 25] investigated dynamic networks gov-
erned by decentralized/descriptor systems, respectively. The
related research has been widely used in the research of
multiagent robot and four-rotor aircraft [26].
Decentralized/interconnected descriptor systems also
contain impulse terms. However, to the best of our knowl-
edge, the research on decentralized impulse controllabil-
ity/observability of interconnected descriptor systems still
receives little attention. So in this paper, we study the problem
of decentralized impulsive controllability/observability for
interconnected descriptor systems. The derivative coefficient
matrix (𝐸 + 𝐵𝐾) is not required to be revertible, so our
results include previous related ones as special cases. Decen-
tralized impulsive controllability/observability is investigated
by using matrix rank properties, which is a novel way to
deal with the impulse. The dynamic order of the closed-loop
descriptor systems can be assigned between the minimum
rank and maximum rank of (𝐸 + 𝐵𝐾).
Hypersonic technology is the best stratagem in the realm
of aerospace technology in the 21st century. The hypersonic
vehicle is characterized by its high speed and high ability
of penetration. So it is of great importance in both civilian
field and military field. Because of the advanced aerodynamic
configuration, fuselage-engine integrated design, large flight
envelope, and the tremendous changes of flight environment,
the hypersonic aircraft has poor stability, serious couplings
among the subsystems, and biggish model uncertainty. Mean-
while, various restrictions for the flight of the vehicles need
to be considered. That means that system equations of the
vehicles should include the algebraic equations. In most cases,
the parameters of hypersonic technology systems change
all the time. When derivative coefficient matrix is singular
matrix, we can describe the hypersonic technology systems
as descriptor systems. So, the impulse problem of vehicle
systems is the first problem to be solved. However, to the
best of our knowledge, the research on impulse control of
hypersonic vehicle still receives little attention. Therefore,
the research problem of hypersonic technology impulse
controllability is overwhelmingly crucial.
As is known to all, for the descriptor systems, propor-
tional feedback can not alter the impulse controllability of
systems. Therefore, this paper focuses on designing a deriva-
tive feedback controller for the first subsystem of the large-
scale interconnected descriptor system such that the second
subsystem is impulse controllable or impulse observable. The
impulse controllability/observability problem of descriptor
systems involves the rank constraints of matrix pencil with
parameter 𝐾. Firstly, we use matrix theory to transform the
matrix pencil involving parameter 𝐾 into the form of 𝐴 +
𝐵𝐾𝐶. Further by comparing the maximum rank of the left
hand of the condition of impulse controllability/observability
with the minimum rank of the right hand of the condition
of impulse controllability/observability, necessary conditions
for the existence of a derivative feedback controller for the
first subsystem ensuring the second subsystem to be impulse
controllable and impulse observable are derived, respectively.
If the maximum rank of the left hand of the condition of
impulse controllability is lower than the minimum rank of
the right hand of that, the controller to be designed will
not exist. If the maximum rank of the left hand of the
condition of impulse controllability/observability is higher
than the minimum rank of the right hand of that, the
intersection of the set of the ranks of the left hand and the
set of the ranks of right hand of the condition of impulse
controllability/observability is not an empty set, and the
controller to be designed will exist and can be selected in
the intersection. Finally, illustrative examples are provided to
demonstrate the applicability of the proposed method.
This paper is structured as follows. In Section 2, defini-
tions and lemmas are introduced for later use. Section 3 gives
conditions of the impulsive controllability and impulsive
observability by derivative feedback. In Section 4, illustrative
examples are proposed to testify the feasibility of the theo-
rems. Finally, conclusive remarks are made in Section 5.
2. Preliminaries
Consider a class of large-scale interconnected descriptor
systems composed of two subsystems of the following form:
𝐸1 𝑥1̇ = 𝐴 11 𝑥1 + 𝐴 12 𝑥2 + 𝐵1 𝑢1 ,
𝐸2 𝑥2̇ = 𝐴 21 𝑥1 + 𝐴 22 𝑥2 + 𝐵2 𝑢2 ,
(1)
𝑦1 = 𝐶1 𝑥1 ,
𝑦2 = 𝐶2 𝑥2 ,
Mathematical Problems in Engineering 3

where 𝑥1 ∈ 𝑅𝑛1 is the state of the first subsystem, 𝑥2 ∈ 𝑅𝑛2 where

is the state of the second subsystem, 𝑦1 ∈ 𝑅𝑚1 and 𝑦2 ∈ 𝐸1 + 𝐵1 𝐾1 0
𝑅𝑚2 are outputs, and 𝑢1 ∈ 𝑅𝑙1 and 𝑢2 ∈ 𝑅𝑙2 are control 𝐸𝐾 = [ ],
inputs. 𝐸1 , 𝐸2 , 𝐴 11 , 𝐴 12 , 𝐴 21 , 𝐴 22 , 𝐵1 , 𝐵2 , 𝐶1 , and 𝐶2 are 0 𝐸2
(7)
constant matrices with compatible dimensions. In this paper, 0
we suppose that 𝐵2 = [ ] .
𝐵2
rank (𝐸1 ) = 𝑟𝑒1 ,
Setting
rank (𝐵1 ) = 𝑟𝑏1 ,
𝐵1 0
𝐵1 = [ ], (8)
rank (𝐸2 ) = 𝑟𝑒2 , 0 0
rank (𝐵2 ) = 𝑟𝑏2 , (2) then we have
𝐸𝐾 = 𝐸 + 𝐵1 𝐾1 . (9)
rank ([𝐸1 𝐵1 ]) = 𝑟𝑒1 𝑏1 ,
The local derivative output feedback can be described as
rank ([𝐸2 𝐵2 ]) = 𝑟𝑒2 𝑏2 ,
𝑢1 = −𝐿 1 𝑦1̇ , (10)
𝑛1 + 𝑛2 = 𝑛.
where 𝐿 1 ∈ 𝑅𝑙1 ×𝑚1 is a gain matrix to be determined. Applying
Denote the feedback to the first subsystem of system (1), the resultant
𝐸1 0 closed-loop system is
𝐸=[ ],
0 𝐸2 𝐸𝐿 𝑥̇ = 𝐴𝑥 + 𝐵2 𝑢2 , (11)
where
𝐴 11 𝐴 12
𝐴=[ ], 𝐸1 + 𝐵1 𝐿 1 𝐶1 0
𝐴 21 𝐴 22 𝐸𝐿 = [ ],
0 𝐸2 (12)
𝐵1 0
𝐵=[ ], 𝐶2 = [0 𝐶2 ] .
0 𝐵2
Setting
𝐶1 0
𝐶=[ ], (3) 𝐶1 0
0 𝐶2 𝐶1 = [ ], (13)
0 0
𝑥1
𝑥 = [ ], then we have
𝑥2
𝐸𝐿 = 𝐸 + 𝐵1 𝐿 1 𝐶1 . (14)
𝑢1 For system (1), there exist orthogonal matrices
𝑢 = [ ],
𝑢2 𝑄1 , 𝑄2 , 𝑈1 , 𝑈2 , 𝑉1 , and 𝑉2 with appropriate dimensions
such that
𝑦1
𝑦 = [ ]. Σ11 0 0
𝑦2 [ 1 ]
𝑈1 𝐸1 𝑉1 = [ 1 ]
[𝐸21 𝐸22 0] ,
Then system (1) can be rewritten into the condensed form
[ 0 0 0]
𝐸𝑥̇ = 𝐴𝑥 + 𝐵𝑢,
(4) 0 0
[ 1 ]
𝑦 = 𝐶𝑥, 𝑈1 𝐵1 𝑄1 = [ ]
[Σ𝐵 0] ,
and system (1) is assumed to be regular; that is, ∃𝑠 ∈
[ 0 0]
C, det(𝑠𝐸 − 𝐴) ≠ 0. In this case, the state response of system (15)
(1) exists uniquely for any admissible initial state. Σ21 0 0
The local derivative feedback can be described as [ 2 ]
𝑈2 𝐸2 𝑉2 = [ 2
[𝐸21 𝐸22 0]
],
𝑢1 = −𝐾1 𝑥1̇ , (5)
[ 0 0 0]
where 𝐾1 ∈ 𝑅𝑙1 ×𝑛1 is a gain matrix to be determined. Applying
0 0
the feedback to the first subsystem of system (1), the resultant [ 2 ]
closed-loop system is 𝑈2 𝐵2 𝑄2 = [ ]
[Σ𝐵 0] ,
𝐸𝐾 𝑥̇ = 𝐴𝑥 + 𝐵2 𝑢2 , (6) [ 0 0]
4 Mathematical Problems in Engineering

where Σ11 ∈ 𝑅(𝑟𝑒1 𝑏1 −𝑟𝑏1 )×(𝑟𝑒1 𝑏1 −𝑟𝑏1) , Σ1𝐵 ∈ 𝑅𝑟𝑏1 ×𝑟𝑏1 , Σ21 ∈ Lemma 3 (see [2, 21]). For arbitrary matrices A ∈ 𝑅𝑛×𝑟 , B ∈
𝑅(𝑟𝑒2 𝑏2 −𝑟𝑏2 )×(𝑟𝑒2 𝑏2 −𝑟𝑏2 ) , and Σ2𝐵 ∈ 𝑅𝑟𝑏2 ×𝑟𝑏2 are diagonal pos- 𝑅𝑛×𝑚 , and C ∈ 𝑅𝑙×𝑟 the following relation holds:
1
itive definite matrices. 𝐸21 ∈ 𝑅𝑟𝑏1 ×(𝑟𝑒1 +𝑟𝑏1 −𝑟𝑒1 𝑏1 ) , 𝐸22
1
∈
𝑟𝑏1 ×(𝑟𝑒1 +𝑟𝑏1 −𝑟𝑒1 𝑏1 ) 2 𝑟𝑏2 ×(𝑟𝑒2 +𝑟𝑏2 −𝑟𝑒2 𝑏2 ) 2
(a) min {rank (A + BK)}
𝑅 , 𝐸21 ∈ 𝑅 , and 𝐸22 ∈ K
(22)
𝑅𝑟𝑏2 ×(𝑟𝑒2 +𝑟𝑏2 −𝑟𝑒2 𝑏2 ) are full column rank matrices: = rank ([A B]) − rank (B) ,
𝐴11 11 11
11 𝐴 12 𝐴 13
[ 11 11 11 ] where minK {∗} is the minimum rank of matrix pencil {∗} with
𝑈1 𝐴 11 𝑉1 = [ ]
[𝐴 21 𝐴 22 𝐴 23 ] , parameter matrix K ∈ 𝑅𝑚×𝑟 .
11 11 11 Consider the following:
[𝐴 31 𝐴 32 𝐴 33 ]
(b) min {rank (A + BKC)}
𝐴12 12 12
11 𝐴 12 𝐴 13
K
[ 12 12 12 ]
𝑈1 𝐴 12 𝑉2 = [ ]
[𝐴 21 𝐴 22 𝐴 23 ] , A
12 12 12
= rank ([A B]) + rank ([ ])
[𝐴 31 𝐴 32 𝐴 33 ] C (23)

𝐴21 21 21
11 𝐴 12 𝐴 13
A B
[ 21 21 21 ] (16) − rank ([ ]) ,
𝑈2 𝐴 21 𝑉1 = [ ] C 0
[𝐴 21 𝐴 22 𝐴 23 ] ,
21 21 21 where minK {∗} is the minimum rank of matrix pencil {∗} with
[𝐴 31 𝐴 32 𝐴 33 ]
parameter matrix K ∈ 𝑅𝑚×𝑙 .
𝐴22 22 22
11 𝐴 12 𝐴 13 Consider the following:
[ 22 22 22 ]
𝑈2 𝐴 22 𝑉2 = [ ]
[𝐴 21 𝐴 22 𝐴 23 ] , (c) max {rank (A + BKC)}
22 22 22 K
[𝐴 31 𝐴 32 𝐴 33 ] (24)
A
1
𝑈1 𝐶1 𝑉1 = [𝐶11 1
𝐶12 1
𝐶13 ], = min {rank ([A B]) , rank ([ ])} ,
C
2 2 2
𝑈2 𝐶2 𝑉2 = [𝐶11 𝐶11 𝐶11 ]. where maxK {∗} is the maximum rank of matrix pencil {∗} with
For linear descriptor systems of the form parameter matrix K ∈ 𝑅𝑚×𝑙 .
Consider the following:
𝐸𝑥̇ = 𝐴𝑥 + 𝐵𝑢,
(17)
𝑦 = 𝐶𝑥, 𝑀 = {K ∈ 𝑅𝑚×𝑙 | rank (A + BKC)
(25)
where 𝑥 ∈ 𝑅𝑛 is state vector, 𝑢 ∈ 𝑅𝑟 is control input, and 𝑦 ∈
𝑅𝑙 is controlled output, 𝐸, 𝐴 ∈ 𝑅𝑛×𝑛 , 𝐵 ∈ 𝑅𝑛×𝑟 , and 𝐶 ∈ 𝑅𝑙×𝑛 = max {rank (A + BKC)}} .
K
are constant matrices and rank(𝐸) = 𝑛0 ≤ 𝑛.
It is a Zariski open set, or equivalently almost every matrix
Definition 1 (see [6]). System (17) is said to be regular if there K ∈ 𝑅𝑚×𝑙 can make rank(A + BKC) to be maximal.
exists a constant scalar 𝑠 ∈ C such that
By Lemma 2, we have the following results.
det (𝑠𝐸 − 𝐴) ≠ 0. (18)
Lemma 2 (see [6]). (a) System (17) is impulse-free if and only Lemma 4. (a) System (6) is impulse controllable if and only if
if the following relation holds: the following relation holds:
𝐸 0 𝐸𝐾 0 0
rank ([ ]) = 𝑛 + rank (𝐸) . (19) rank ([ ]) = 𝑛 + rank (𝐸𝐾 ) . (26)
𝐴 𝐸 𝐴 𝐸𝐾 𝐵2
(b) System (17) is impulse controllable if and only if the (b) System (11) is impulse observable if and only if the
following relation holds: following relation holds:
𝐸 0 0
rank ([ ]) = 𝑛 + rank (𝐸) . (20) 𝐸𝐿 𝐴
𝐴 𝐸 𝐵 [ ]
rank ([ 0 𝐸𝐿 ]
[
]) = 𝑛 + rank (𝐸𝐿 ) . (27)
(c) System (17) is impulse observable if and only if the
following relation holds: [ 0 𝐶2 ]
𝐸 𝐴 Lemma 5 (see [29]). If 𝑟min = minK rank(A + BKC) and
[ 0 𝐸] 𝑟max = maxK rank(A + BKC), then there always exists K ∈
rank ([ ]) = 𝑛 + rank (𝐸) . (21)
𝑅𝑚×𝑙 such that 𝑟0 = rank(A + BKC) for any positive integer
[ 0 𝐶] 𝑟0 satisfying 𝑟min ≤ 𝑟0 ≤ 𝑟max .
Mathematical Problems in Engineering 5

3. Main Results = 2 rank (Σ11 ) + 2 rank (Σ1𝐵 ) + 2 rank (Σ21 )

2
In this section, conditions for the existence of a derivative + rank (𝐸22 ) + rank (Σ2𝐵 )
feedback controller for the first subsystem of the large-
scale interconnected descriptor systems ensuring the second 𝐴11 11 12
32 𝐴 33 𝐴 33
subsystem to be impulse controllable and impulse observable + rank ([ 21 21 11 ])
are derived, respectively. 𝐴 32 𝐴 33 𝐴 33

Theorem 6. Consider system (1). If 𝐴11 11 12

32 𝐴 33 𝐴 33
= 2𝑟𝑒1 𝑏1 + 𝑟𝑒2 + 𝑟𝑒2 𝑏2 + rank ([ 21 21 22 ]) .
𝐴 32 𝐴 33 𝐴 33
𝐴11 11 12
32 𝐴 33 𝐴 33 (32)
rank ([ 21 21 22 ]) < 𝑛 − 𝑟𝑒1 𝑏1 − 𝑟𝑒2 𝑏2 − 𝑟𝑏1 , (28)
𝐴 32 𝐴 33 𝐴 33
On the other hand, by Lemma 3, we have
there does not exist 𝐾1 such that closed-loop system (6) is
𝑟1 = 𝑛 + rank (𝐸𝐾 )
impulse controllable.
𝐸1 + 𝐵1 𝐾1 0
Proof . We first note that = min {𝑛 + rank ([ ])}
𝐾1 0 𝐸2
(33)
𝐸𝐾 0 0 𝐸 0 0
[ ]=[ ] 𝐸1 0 𝐵1 0 𝐵1 0
𝐴 𝐸𝐾 𝐵2 𝐴 𝐸 𝐵2 = 𝑛 + rank ([ ]) − rank ([ ])
0 𝐸2 0 0 0 0
𝐼 0 0 (29)
𝐵1 0 0 = 𝑛 + 𝑟𝑒2 + 𝑟𝑒1 𝑏1 − 𝑟𝑏1 .
[0 𝐼 0 ]
+[ ] 𝐾1 [ ],
0 𝐵1 0 When 𝑅1 < 𝑟1 , namely, (28) holds, by Lemma 4, we
[0 0 𝐼 ] can know that there does not exist 𝐾1 such that closed-
loop system (6) is impulse controllable. This completes the
where proof.
̃1
𝐾 0 0 Theorem 7. Consider system (1). If
[ ]
𝐾1 = [
[0 ̃1 0]
𝐾 ], 𝐴11 11 12
32 𝐴 33 𝐴 33
rank ([ 21 21 22 ]) ≥ 𝑛 − 𝑟𝑒1 𝑏1 − 𝑟𝑒2 𝑏2 − 𝑟𝑏1 ,
[0 0 0] (30)
𝐴 32 𝐴 33 𝐴 33
(34)
𝐾1 0
̃1 = [
𝐾 ]. Ω𝐾 = ⋃ {𝐾1𝑖 ∩ 𝐾2𝑖 } ≠ 𝜙,
0 𝐾1 𝑟1 ≤𝑖≤𝑅1

Denote where

S1 = Diag {𝑈1 , 𝑈2 , 𝑈1 , 𝑈2 } , 𝐸𝐾 0 0
𝐾1𝑖 = {𝐾1 | rank ([ ]) = 𝑖} ,
(31) 𝐴 𝐸𝐾 𝐵 2
S2 = Diag {𝑉1 , 𝑉2 , 𝑉1 , 𝑉2 , 𝑄2 , 𝑄1 , 𝑄1 } . (35)
𝐾2𝑖 = {𝐾1 | 𝑛 + rank (𝐸𝐾 ) = 𝑖} ,
Diag{∗} denotes the block diagonal matrix. By Lemma 3, we
have then one can find 𝐾1 ∈ Ω𝐾 such that the closed-loop system is
impulse controllable.
𝐸𝐾 0 0
𝑅1 = max {rank ([ ])} Proof. The proof of Theorem 7 is similar to that of Theorem 6;
𝐾1 𝐴 𝐸𝐾 𝐵 2 the details are omitted.
𝐸 0 0 𝐵1 0 Remark 8. In the case of 𝑅1 < 𝑟1 , there does not exist
= max {rank ([ ])}
𝐾1 𝐴 𝐸 𝐵2 0 𝐵1 gain matrix 𝐾1 such that the closed-loop system is impulse
controllable. When 𝑅1 ≥ 𝑟1 , the existence of gain matrix 𝐾1
𝐸1 0 0 0 0 𝐵1 0 can be considered in terms of probability. Namely, assume
[ ] there exists a topological space, in which we can define a
[ 0 𝐸2 0 0 0 0 0]
= rank (S1 [ ] S2 ) ̃ that expresses the intersection of the dense set
nonnull set Ω
[𝐴 𝐴 𝐸 0 0 0 𝐵1 ]
[ 11 12 1 ] and any subset of that topological space. Then based on the
[𝐴 21 𝐴 22 0 𝐸2 𝐵2 0 0] ̃
perspective of probability, we can easily obtain this kind of Ω.
6 Mathematical Problems in Engineering

Based on this idea, we know that 𝐾1𝑅1 is a Zariski open Theorem 10. Consider system (1). If
set according to Lemma 3. We also know that almost every
𝐾1 can make rank([ 𝐸𝐴𝐾 𝐸0 𝐵0 ]) = 𝑅1 . Namely, 𝐾1𝑅1 is a dense 𝐸 𝐴 𝐵1 0
𝐾 2
[ ]
set. According to Lemma 5 we know that there always exists rank ([ ]
[ 0 𝐸 0 𝐵1 ])
𝐾1 such that 𝑛 + rank(𝐸𝐾 ) = 𝑅1 , and 𝐾2𝑅1 is a subset of
the topological space 𝑅𝑚×𝑛1 . Then based on the perspective [ 0 𝐶2 0 0 ]
̃ 𝐾 . If Ω
of probability, we can easily get this kind of Ω ̃ 𝐾 ≠ 0, we
𝐸 𝐴
̃
can find 𝐾1 ∈ Ω𝐾 such that the closed-loop system is impulse [ ]
[0 𝐸]
controllable, where [ ]
([ ])
̃ 𝐾 = 𝐾1𝑅 ∩ 𝐾2𝑅 . ≤ rank ( [
([ 0 𝐶2 ])
]) ,
Ω (36) [ ]
1 1
[𝐶 0]
[ 1 ] (42)
Remark 9. When Ω ̃ 𝐾 ≠ 0 and 𝑛 + rank(𝐸𝐾 ) = 𝑅1 , we can
obtain 𝐾1 as the following form: ([ 0 𝐶1 ])

−1
𝐴132 𝐴133 𝐴12
33
𝐾𝑑11 [− (Σ1𝐵 ) 1
𝐸22 0] + Δ [ 21 21 2 ]
𝐾1 = 𝑄[ ] 𝑉𝑇 , (37) [𝐴 𝐴 𝐴 ]
1
[ 22 23 23 ]
𝐾𝑑21 [𝐾𝑑22 𝐾𝑑23 ] rank ([ 21 21 2 ]) ≥ 𝑛 − 𝑟𝑒1 𝑏1 − 𝑟𝑏1 − 𝑟𝑒2 ,
[𝐴 𝐴 𝐴 ]
[ 32 33 33 ]
where 𝐾𝑑11 ∈ 𝑅𝑟𝑏 ×(𝑟𝑒1 𝑏1 −𝑟𝑏1) , 𝐾𝑑21 ∈ 𝑅(𝑙1 −𝑟𝑏1 )×(𝑟𝑒1 𝑏1 −𝑟𝑏1 ) , 𝐾𝑑22 ∈ [ 0 0 𝐶132
]
𝑅(𝑙1 −𝑟𝑏1 )×(𝑟𝑒1 +𝑟𝑏1 −𝑟𝑒1 𝑏1) , and 𝐾𝑑23 ∈ 𝑅(𝑙1 −𝑟𝑏1 )×(𝑛−𝑟𝑒1 ) are four
arbitrary parameters matrices, Δ ∈ 𝑅𝑟𝑏1 ×(𝑛−𝑟𝑒1 𝑏1 +𝑟𝑏1 ) is full Ω𝐿 = ⋃ {𝐿 1𝑖 ∩ 𝐿 2𝑖 } ≠ 𝜙,
𝑟2 ≤𝑖≤𝑅2
column rank parameter matrix, and rank(Δ) = 𝑅1 − 𝑛 − 𝑟𝑒1 𝑏1 +
𝑟𝑏1 −𝑟𝑒2 . At the same time, if 𝐾1 which is obtained by the above
process satisfies formula (27), then 𝐾1 is the gain matrix. Let where

𝐾𝑑11 𝐾𝑑12 𝐾𝑑13 { 𝐸𝐿 𝐴 }

𝑄1 𝑇 𝐾1 𝑉1 = [ ], (38) {
{ [ ] }
}
𝐾𝑑21 𝐾𝑑22 𝐾𝑑23 [ ]
𝐿 1𝑖 = {𝐿 | rank ([ 0 𝐸𝐿 ]) = 𝑖} ,
{
{ }
} (43)
in which matrix-blocks are compatible. Then we have { [ 0 𝐶 2] }

𝑈𝐸𝐾 𝑉 𝐿 2𝑖 = {𝐿 | 𝑛 + rank (𝐸𝐿 ) = 𝑖} ,

Σ11 0 0 0 0 0 then one can find 𝐿 ∈ Ω𝐿 such that closed-loop system (11) is
[ 1 1 1 1 1 ]
[𝐸21 + Σ1 𝐾𝑑11 𝐸22 + Σ𝐵 𝐾𝑑12 Σ𝐵 𝐾𝑑13 0 0 0] impulse observable.
[ 1 1 ]
[ ]
[ 0 0 0 0 0 0] (39)
=[
[
],
]
Proof. Note that
[ 0 0 0 Σ21 0 0]
[ ]
[ 2 2 ]
[ 0 0 0 𝐸21 𝐸22 0] 𝐸𝐿 𝐴 𝐸 𝐴 𝐵1 0
[ ] [ ] [ ] 𝐶 0
[ 0 0 0 0 0 0] [ 0 𝐸 ] = [ 0 𝐸 ] + [ 0 𝐵 ] 𝐿1 [ 1 ], (44)
[ 𝐿] [ ] [ 1]
0 𝐶1
where 𝑈 = [ 𝑈1 𝑈2 ], 𝑉 = [ 𝑉1 𝑉2 ]. [ 0 𝐶2 ] [ 0 𝐶2 ] [ 0 0 ]
By rank(𝐸𝐾 ) = 𝑅1 − 𝑛 and (39) we can get the following
formula: where
1
rank ([𝐸22 + Σ1𝐵1 𝐾𝑑12 Σ1𝐵1 𝐾𝑑13 ]) 𝐿̃1 0
(40) 𝐿1 = [ ],
0 𝐿̃1
= 𝑅1 − 𝑛 − 𝑟𝑒1 𝑏1 + 𝑟𝑏1 − 𝑟𝑒2 .
(45)
𝐿1 0
Obviously, the matrix [𝐾𝑑12 𝐾𝑑13 ] can be expressed as the ̃1 = [
𝐿 ].
following form: 0 𝐿1

−1
[𝐾𝑑12 𝐾𝑑13 ] = [− (Σ1𝐵 ) 𝐸22
1
0] + Δ, Denote
1
(41)
rank (Δ) = 𝑅1 − 𝑛 − 𝑟𝑒1 𝑏1 + 𝑟𝑏1 − 𝑟𝑒2 . S3 = Diag {𝑈1 , 𝑈2 , 𝑈1 , 𝑈2 , 𝑈1 , 𝑈2 } ,
(46)
Next, we discuss the problem of impulse observability. S4 = Diag {𝑉1 , 𝑉2 , 𝑉1 , 𝑉2 , 𝑄1 , 𝑄1 } ,
Mathematical Problems in Engineering 7

when Since 𝑅2 > 𝑟2 and Ω𝐿 ≠ 𝜙, by Lemma 4 there exists 𝐿 1

such that closed-loop system (11) is impulse observable. This
𝐸 𝐴 𝐵1 0 completes the proof.
[ ]
rank ([ 0 𝐸 0 𝐵1 ]
[
]) Theorem 11. Consider system (1). If
[ 0 𝐶2 0 0 ]
𝐸 𝐴 𝐵1 0
𝐸 𝐴 [ ]
(47) rank ([ ]
[ 0 𝐸 0 𝐵1 ])
[ ]
[0 𝐸]
[ ] [ 0 𝐶2 0 0 ]
([ ])
(
≤ rank ([[0 𝐶2 ] )
]) .
[ ] 𝐸 𝐴
[𝐶 0] [ ]
[ 1 ] [0 𝐸]
[ ]
([ 0 𝐶 1 ]) ([ ])
> rank ( [
([ 0 𝐶] )
]) ,
[ ]
By Lemma 3, (15), and (16), we have [𝐶 0]
[ 1 ]
{
{
𝐸𝐿 𝐴 } ([ 0 𝐶1 ])
{ [ ] }
}
[ ]
𝑅2 = max {rank ([ 0 𝐸𝐿 ])}
𝐿1 { } 0 𝐴123 𝐴12
23 (50)
{ } [ ]
{ [ 0 𝐶2] } [ 0 𝐴133 𝐴12 ]
[ 33 ]
([ ])
𝐸 𝐴 𝐵1 0 ([ 0 𝐴21 𝐴223 ] )
([ 23 ])
[ ] ([ ] )
= rank ([ ] rank ([
([ 0 𝐴21 𝐴233 ] )
[ 0 𝐸 0 𝐵1 ]) ([
33 ])
])
([ 0 0 2 ])
𝐶13 ])
[ 0 𝐶2 0 0 ] ([
[ 1 ]
[𝐶 0 0 ]
𝐸1 0 𝐴 1 𝐴 12 𝐵1 0 [ 13 ]
[ ] 1
[ 0 𝐸2 𝐴 21 𝐴 2 0 0] ([ 0 𝐶13 0 ])
[ ]
([ 0 0 𝐸 0 0 𝐵1 ] ) ≥ 𝑛 + 𝑟𝑒1 𝑏1 − 2𝑟𝑒1 − 𝑟𝑒2 − 𝑟𝑏1 ,
= rank ([ 1 ])
[ ]
[0 0 0 𝐸 0 0]
[ 2 ] Ω𝐿 = ⋃ {𝐿 1𝑖 ∩ 𝐿 2𝑖 } ≠ 𝜙,
(48)
([ 0 0 0 𝐶2 0 0 ]) 𝑟2 ≤𝑖≤𝑅2

𝐸1 0 𝐴 11 𝐴 12 𝐵1 0 then one can find 𝐿 ∈ Ω𝐿 such that closed-loop system (11) is

[ ] impulse observable.
[0 𝐸2 𝐴 21 𝐴 22 0 0]
[ ]
( [0 0 𝐸1 0 0 𝐵1 ] ) Proof. The proof of Theorem 11 is similar to that of
= rank (S3 [ ] S4 )
[ ] Theorem 10; the details are omitted.
[0 0 0 𝐸2 0 0]
[ ]
( [0 0 0 𝐶2 0 0] ) 4. Illustrative Examples
𝐴132 𝐴133 𝐴12
33 In this section, illustrative examples are considered to show
[ 21 21 2 ] the effectiveness of the proposed approach.
[𝐴 𝐴 𝐴 ]
[ 22 23 23 ]
= 2𝑟𝑒1 𝑏1 + 2𝑟𝑒2 + rank ([ 21 21 2 ]) .
[𝐴 𝐴 𝐴 ] Example 1. Consider system (1) with the following parame-
[ 32 33 33 ]
2 ters:
[ 0 0 𝐶13 ]
1 3
On the other hand, by Lemma 3, we have 𝐸1 = [ ],
2 6

𝑟2 = min {𝑛 + rank (𝐸𝐿 )} −2.3 4.5

𝐿1 𝐸2 = [ ],
1.15 −2.25
= 𝑛 + rank ([𝐸 𝐵1 ]) − rank (𝐵1 )
−1 0.5
(49) 𝐴 11 = [ ],
𝐸1 0 𝐵1 0 𝐵1 0 −2 1
= 𝑛 + rank ([ ]) − rank ([ ])
0 𝐸2 0 0 0 0
1 6
= 𝑛 + 𝑟𝑒2 + 𝑟𝑒1 𝑏1 − 𝑟𝑏1 . 𝐴 22 = [ ],
−0.5 −3
8 Mathematical Problems in Engineering

4 0.7 𝑟1 = min {𝑛 + rank (𝐸𝐾 )}

𝐴 12 = [ ], 𝐾1
8 1.4
𝐸1 + 𝐵1 𝐾1 0
−3 1.5 = min {𝑛 + rank ([ ])}
𝐴 21 = [ ], 𝐾1 0 𝐸2
1.5 −0.75
𝐸1 0 𝐵1 0
0.5 = 𝑛 + rank ([ ])
𝐵1 = [ ], 0 𝐸2 0 0
1
𝐵1 0
2 − rank ([ ])
𝐵2 = [ ]. 0 0
−1
(51) 1 3 0 0 0.5
[ ]
[2 6 0 0 1]
= 4 + rank ([
[0 0 −2.3
])
[ 4.5 0]
]
Note that
[0 0 1.15 −2.25 0 ]
0.5
𝐸𝐾 0 0 − rank ([ ]) = 5.
𝑅1 = max {rank ([ ])} 1
𝐾1 𝐴 𝐸𝐾 𝐵2 (52)
𝐸1 0 0 0 0 𝐵1 0 𝑅1 < 𝑟1 does not meet the necessary condition of impulse
[ ]
[ 0 𝐸2 0 0 0 0 0] controllability.
= rank ([
[𝐴 𝐴 𝐸 0
]) By Theorem 6, there does not exist 𝐾1 such that closed-
[ 11 12 1 0 0 𝐵1 ]
] loop system (6) is impulse controllable.
[𝐴 21 𝐴 22 0 𝐸2 𝐵2 0 0] We can design a derivative feedback controller for the first
subsystem as follows:
2 6 0 0
[0 0 −2.3 4.5] 𝑢1 = −𝐾1 𝑥1̇ , 𝐾1 = [𝑎 𝑏] , (53)
= 1 + rank ([ ]) = 4,
[−1 0.5 4 0.7] where 𝑎, 𝑏 are arbitrary parameters. We have

𝐸1 + 𝐵1 𝐾1 0 0 0 0
[ ]
𝐸𝐾 0 0 [ 0 𝐸2 0 0 0]
rank ([ ]) = rank ([
[
])
𝐴 𝐸𝐾 𝐵2 [ 𝐴 11 𝐴 12 𝐸1 + 𝐵1 𝐾1 0 0]
]
[ 𝐴 21 𝐴 22 0 𝐸2 𝐵2 ]
𝑎 𝑏
+1 +3 0 0 0 0 0 0 0
[2 ]
[ 𝑎 + 2 2𝑏 + 6 0 0 0 0 0 0 0]
[ ]
([ ])
([ 0 0 −2.3 4.5 0 0 0 0 0] )
([ ])
([ ])
([ 0 0 1.15 −2.25 0 0 0 0 0] )
([ ])
= rank ([ ]) ≤ 5, (54)
([ 𝑎 𝑏 ])
([ −1 0.5 4 0.7 +1 +3 0 0 0] )
([ 2 2 ])
([ ])
([ −2 1 8 1.4 𝑎 + 2 𝑏 + 6 0 0 0 ])
[ ]
[ ]
[ −3 1.5 1 6 0 0 −2.3 4.5 2]
([ −1.5 0.75 −0.5 −3 0 0 1.15 −2.25 −1])
𝑎 𝑏
+1 +3 0 0
[2 2 ]
𝐸1 + 𝐵1 𝐾1 0 [ ]
[𝑎 + 2 𝑏 + 6 0 0 ]
𝑛 + rank (𝐸𝐾 ) = 𝑛 + rank ([ ]) = 4 + rank ([ ]) ≥ 6.
0 𝐸2 [ 0 0 −2.3 4.5 ]
[ ]
[ 0 0 1.15 −2.25]
Mathematical Problems in Engineering 9

For any derivative feedback gain 𝐾1 , equality (26) in Lemma 4 𝑅1 > 𝑟1 satisfies the necessary condition of impulse
can not hold. That means there does not exist 𝐾1 such that controllability. By Theorem 7, we can design proportional
closed-loop system (6) is impulse controllable. feedback controller for the first subsystem as follows:

Example 2. Consider system (1) with the following parame-

ters: 𝑢1 = 𝐾𝑑 𝑥, 𝐾𝑑 = [𝑎 𝑏] . (57)
1 0
𝐸1 = [ ],
0 0
We have
2 0
𝐸2 = [ ],
0 0
𝐸 0 0
rank ([ ])
1 2 𝐴𝐾 𝐸 𝐵2
𝐴 11 =[ ],
5 6
𝐸1 0 0 0 0
1 3 [ ]
[ 0 𝐸2 0 0 0]
𝐴 22 =[ ], = rank ([ ])
2 4 [𝐴 + 𝐵 𝐾 𝐴 𝐸 0 0]
[ 11 1 𝑑 12 1 ]
(55)
7 3 [ 𝐴 21 𝐴 22 0 𝐸2 𝐵2 ]
𝐴 12 =[ ],
4 2 1 0 0 0 0 0 0 0 0
[ ]
2 1 [ 0 0 0 0 0 0 0 0 0]
𝐴 21 = [ ], [ ]
([ ])
4 6 ([ 0 0 2 0 0 0 0 0 0])
([ ])
([ ])
1 ([ 0 0 0 0 0 0 0 0 0])
= rank ( [
([ 1 + 𝑎 2 + 𝑏
])
𝐵1 = [ ] ,
0 ([ 7 3 1 0 0 0 0] )
]) (58)
([ ])
([ 5 6 2 4 0 0 0 0 0] )
([ ])
3 [ ]
𝐵2 = [ ] . [ 2 1 1 3 0 0 2 0 3]
[ ]
0
Note that ([ 4 6 2 4 0 0 0 0 0])

𝐸𝐾 0 0 = 5,
𝑅1 = max {rank ([ ])}
𝐾1 𝐴 𝐸𝐾 𝐵2 1 0 0 0
[ ]
𝐸1 0 0 0 0 𝐵1 0 [0 0 0 0]
𝑛 + rank (𝐸) = 4 + rank ([ ]
[0 0 2 0]) = 6,
[ ] [ ]
[ 0 𝐸2 0 0 0 0 0]
= rank ([
[𝐴 𝐴 𝐸 0
])
[ 11 12 1 0 0 𝐵1 ]
] [0 0 0 0]

[𝐴 21 𝐴 22 0 𝐸2 𝐵2 0 0] 𝐸 0 0
rank ([ ]) ≠ 𝑛 + rank (𝐸) ,
= 6, 𝐴𝐾 𝐸 𝐵2

𝑟1 = min {𝑛 + rank (𝐸𝐾 )}

𝐾1
(56) where 𝐴𝐾 = [ 𝐴 11𝐴+𝐵211 𝐾𝑑 𝐴
𝐴 22 ] .
12

𝐸1 + 𝐵1 𝐾1 0 We can conclude that there does not exist proportional

= min {𝑛 + rank ([ ])} feedback 𝐾𝑑 such that closed-loop system (6) is impulse
𝐾1 0 𝐸2
controllable.
𝐸1 0 𝐵1 0
= 𝑛 + rank ([ ]) Example 3. Consider system (1) with the parameters as
0 𝐸2 0 0 Example 2.
𝑅1 > 𝑟1 satisfies the necessary condition of impulse con-
𝐵1 0
− rank ([ ]) trollability. Further, by Theorem 7 we can design a derivative
0 0 feedback controller for the first subsystem as follows:
1 0 1
= 4 + rank ([ ]) − rank ([ ]) = 5.
0 2 0 𝑢1 = −𝐾1 𝑥1̇ , 𝐾1 = [𝑎 𝑏] . (59)
10 Mathematical Problems in Engineering

Then we have Example 4. The hypersonic vehicle is modeled as in [27],

and the mathematical model (62) of hypersonic vehicle is
established in [28]. Consider
𝐸𝐾 0 0
rank ([ ]) 𝛾̇ = 𝜔𝑥 − tan 𝜗 (𝜔𝑦 cos 𝛾 − 𝜔𝑧 sin 𝛾) ,
𝐴 𝐸𝐾 𝐵2
𝜔𝑥 𝛽
𝐸1 + 𝐵1 𝐾1 0 0 0 0 𝛼̇ = 𝑏3 𝜔𝑧 + 𝑏4 𝛼 + 𝑏4 𝛼̇ + 𝑏8 𝛿𝑧 − ,
[ ] 57.3
[ 0 𝐸2 0 0 0]
= rank ([
[
])
𝛽 ̇ = 𝑎1 𝜔𝑥 + 𝑎2 𝜔𝑦 + 𝑎5 𝛽 + 𝑎5 𝛽 ̇ + 𝑏6 𝛿𝑥 + 𝑏6 𝛿𝑥 + 𝑎7 𝛿𝑦
[ 𝐴 11 𝐴 12 𝐸1 + 𝐵1 𝐾1 0 0]
]
𝐴 21 𝐴 22 0 𝐸2 𝐵2 ] 𝜔𝑥 𝛼
[ + ,
57.3
1+𝑎 𝑏 0 0 0 0 0 0 0
[ ] 𝜔̇ 𝑥 = 𝑐1 𝜔𝑥 + 𝑐2 𝜔𝑦 + 𝑐4 𝛼 + 𝑐5 𝛽 + 𝑐6 𝛿𝑥 + 𝑐7 𝛿𝑦
[ 0 0 0 0 0 0 0 0 0] (62)
[ ]
([ ]) 𝜔𝜔
([ 0 0 2 0 0 0 0 0 0 ]) + (𝐼𝑦 − 𝐼𝑧 ) 𝑧 𝑥 ,
([ ]) (60) 57.3𝐼𝑥
([ ])
([ 0 0 0 0 0 0 0 0 0 ])
= rank (
([
[ ])
𝜔̇ 𝑦 = 𝑏1 𝜔𝑥 + 𝑏2 𝜔𝑦 + 𝑏5 𝛽 + 𝑏5 𝛽 ̇ + 𝑏6 𝛿𝑥 + 𝑏7 𝛿𝑦
([ 1 2 7 3 1 + 𝑎 𝑏 0 0 0]
])
)
([ ])
([ 5 6 4 2 0 0 0 0 0] ) 𝜔𝑧 𝜔𝑥
([ ]) + (𝐼𝑧 − 𝐼𝑥 ) ,
[ ] 57.3𝐼𝑦
[ 2 1 1 3 0 0 2 0 3]
[ ]
𝜔𝑥 𝜔𝑦
([ 4 6 2 4 0 0 0 0 0 ])
𝜔̇ 𝑧 = 𝑎3 𝜔𝑧 + 𝑎4 𝛼 + 𝑎4 𝛼̇ + 𝑎8 𝛿𝑧 + (𝐼𝑥 − 𝐼𝑦 ) ,
57.3𝐼𝑧
1+𝑎 𝑏 0
[ 5 6 2] where 𝛾 denotes roll angle, 𝛼 denotes angle of attack,
= 3 + rank ([ ]) , 𝛽 denotes yaw angle, and 𝜔𝑥 , 𝜔𝑦 , and 𝜔𝑧 denote rotating
[ 4 6 4] angular velocity components in the coordinates 𝑥, 𝑦, and 𝑧.

𝑛 + rank (𝐸𝐾 ) = 5 + rank ([1 + 𝑎 𝑏]) . Various restrictions for the flight of the hypersonic tech-
nology need to be considered. In most cases, the parameters
of hypersonic technology systems change all the time, with
In fact, by the above derivation, we can give arbitrary derivative coefficient matrix into singular matrix; by [29]
derivative feedback controller such that closed-loop system we describe the hypersonic technology systems as descriptor
(6) is impulse controllable. When 𝐾1 = [1 −1], we have systems. By Assumption 2.1 of [30] if the equilibrium of
nonlinear singular system 𝐸𝑥̇ = 𝑓(𝑥) is 𝑥𝑒 = 0 then
the locally nonimpulsiveness of this system is equivalent
𝐸𝐾 0 0 to (deg(det(𝑠𝐸 − 𝐴)) = rank(𝐸) ∀𝑠 ∈ C, where 𝐴 =
rank ([ ])
𝐴 𝐸𝐾 𝐵2 (𝜕𝑓/𝜕𝑥)(0)). Therefore nonlinear large-scale interconnected
descriptor systems (62) can be linearized at equilibrium point
2 −1 0 0 0 0 0 0 (𝛾, 𝛼, 𝛽, 𝜔𝑥 , 𝜔𝑦 , 𝜔𝑧 ) = (0, 0, 0, 0, 0, 0); then we obtain the
[ ] following system:
[0 0 2 0 0 0 0 0]
[ ]
([ ]) 𝐸1 𝑥1̇ = 𝐴 11 𝑥1 + 𝐴 12 𝑥2 + 𝐵1 𝑢1 ,
([ 1 2 7 3 2 −1 0 0])
= rank ( [
([ 5
]) = 6, (61) (63)
([ 6 4 2 0 0 0 0]])
) 𝐸2 𝑥2̇ = 𝐴 21 𝑥1 + 𝐴 22 𝑥2 + 𝐵2 𝑢2 ,
[ ]
[2 1 1 3 0 0 2 3] where
[ ]
1 0 0
( [ 4 6 2 4 0 0 0 0] )
[0 1 − 𝑏 0 ]
𝐸1 = [ 4 ],
2 −1 0
𝑛 + rank (𝐸𝐾 ) = 4 + rank ([ ]) = 6, [0 −𝑎4 −1]
0 0 2
1 0 0
[0 1 − 𝑎 1]
and we can conclude that closed-loop system (6) is impulse 𝐸2 = [ 5 ],
controllable under the above derivative feedback controller.
[0 −𝑏5 1]
Illustrative Examples 2 and 3 are provided for demon-
strating the different effects of proportional feedback and 1 0 0
derivative feedback. We can find that our method of deriva- [0 𝑏 𝑎 ]
tive feedback can change impulsive controllability when the 𝐴 11 =[ 4 3] ,
method of proportional feedback is unable to do it. [0 𝑎4 𝑎4 ]
Mathematical Problems in Engineering 11

1 0 0 𝑎3 = 2,
[ ]
𝐴 12 =[
[
0 0 0] ,
] 𝑎4 = 0.5,
[ 0 0 0]
𝑎5 = 1,
𝑐1 𝑐5 𝑐2
[ ] 𝑎6 = 0,
𝐴 21 = [ ]
[𝑎1 𝑎5 𝑎2 ] ,
𝑎7 = 0.6,
[ 𝑏1 𝑏5 𝑏2 ]
𝑎8 = 0.5,
0 0 0
[ ] 𝑏1 = 0.3,
𝐴 22 =[ ]
[0 0 0] ,
[0 𝑐4 0] 𝑏2 = 0.4,

0 𝑏4 = 1,
[ ]
𝐵1 = [ ]
[ 𝑏8 ] , 𝑏5 = 1,

[𝑎8 ] 𝑏6 = 0,
𝑏6 𝑏7 𝑏7 = 1,
[ ]
𝐵2 = [ ]
[𝑎6 𝑎7 ] , 𝑏8 = 0.5,
[ 𝑏6 𝑏6 ]
𝑐1 = 0.1,
𝛾
[ ] 𝑐2 = 0.4,
𝑥1 = [ ]
[𝛼] ,
𝑐4 = 0.3,
[𝛽]
𝑐5 = 0.4.
𝜔𝑥 (65)
[ ]
𝑥2 = [ ]
[𝜔𝑦 ] ,
[𝜔𝑧 ]
Note that
𝑢1 = 𝛿𝑧 ,

𝛿𝑥
𝑢2 = [ ] . { 𝐸𝐾 0 0 }
𝛿𝑦 𝑅1 = max {rank ([ ])
𝐾1 }
(64) { [ 𝐴 𝐸𝐾 𝐵2 ] }
System (63) is the large-scale interconnected descriptor 𝐸1 0 0 0 0 𝐵1 0
systems for the conditions that 1 − 𝑏4 = 0, 1 − 𝑎5 = 0, [ ]
and 𝐸1 and 𝐸2 are singular matrices. Furthermore, we find [ 0 𝐸 0 0 0 0 0]
[ 2 ]
rank([ 𝐴𝐸 𝐸0 ]) ≠ 𝑛 + rank(𝐸), so impulse exists in the system. = rank ([
[
])
]
[ 𝐴 1 𝐴 12 𝐸1 0 0 0 𝐵1 ]
For hypersonic vehicle, it may generate impulse response to [ ]
the vehicle with the vehicle flight with acceleration, decelera- ([𝐴 21 𝐴 22 0 𝐸2 𝐵2 0 0 ])
tion, and sharp turns, which may lead to the instability of the
system or may even destroy the vehicle, and hence it is not = 11,
expected to exist.
Consider system (63) with the following parameters [22, 𝑟1 = min {𝑛 + rank (𝐸𝐾 )}
23]: 𝐾1

𝑎1 = 0.5, 𝐸1 + 𝐵1 𝐾1 0
= min {𝑛 + rank ([ ])}
𝑎2 = 0.8, 𝐾1 0 𝐸2
12 Mathematical Problems in Engineering

𝐸1 0 𝐵1 0 𝑛 + rank (𝐸)
= 𝑛 + rank ([ ])
0 𝐸2 0 0
1 0 0 0 0 0
𝐵1 0 [ ]
[0 0 0 0 0 0]
− rank ([ ]) = 10. [ ]
0 0 ([ ])
([0 −0.5 −1 0 0 0])
= 6 + rank ( [
([0 0
]) = 11.
(66)
([ 0 1 0 0]
])
)
[ ]
[0 0 0 0 0 1]
𝑅1 > 𝑟1 satisfies the necessary condition of impulse [ ]
controllability. By Theorem 7, we can design a proportional
([0 0 0 0 −1 1])
feedback controller for the first subsystem as follows:
(68)
𝑢1 = 𝐾𝑑 𝑥, 𝐾𝑑 = [𝑎 𝑏 𝑐] . (67)
Obviously,
Then we have
𝐸 0 0
𝐸 0 0 rank ([ ]) ≠ 𝑛 + rank (𝐸) . (69)
rank ([ ]) 𝐴𝐾 𝐸 𝐵2
𝐴𝐾 𝐸 𝐵2
By the above derivation, for any proportional feedback gain
𝐸1 0 0 0 0 𝐾𝑑 , equality (26) in Lemma 4 can not hold. That means
[ ] there does not exist 𝐾𝑑 such that closed-loop system (6) is
[ 0 𝐸2 0 0 0]
= rank ([
[𝐴 + 𝐵 𝐾 𝐴 𝐸 0
]) impulse controllable. Further, by Theorem 7 we can design
[ 11 1 𝑑 12 1 0]
] a derivative feedback controller for the first subsystem as
𝐴 21 𝐴 22 0 𝐸2 𝐵2 ] follows:
[
−0.5 −1 𝑢1 = −𝐾1 𝑥1̇ , 𝐾1 = [𝑎 𝑏 𝑐] . (70)
= 9 + rank ([ ]) = 10,
1 2 Then we have

𝐸1 + 𝐵1 𝐾1 0 0 0 0
[ ]
𝐸𝐾 0 0 [ 0 𝐸2 0 0 0]
[ ]
rank ([ ]) = rank ([
[
])
]
[ 𝐴 11 𝐴 12 𝐸1 + 𝐵1 𝐾1 0 0 ]
[ 𝐴 𝐸𝐾 𝐵2 ] [ ]
([ 𝐴 21 𝐴 22 0 𝐸2 𝐵2 ])

1 + 0.1𝑎 0.1𝑏 0.1𝑐 0 0 0

[ ]
[ 0 0 0 0 0 0 ]
[ ]
([ ]) (71)
([ 0.2𝑎 −0.5 + 0.2𝑎 −1 + 0.2𝑐 0 0 0 ])
([ ])
= 6 + rank (
([
[ ]) ,
])
([ 1 0 0 1 + 0.1𝑎 0.1𝑏 0.1𝑐 ])
([ ])
[ ]
[ 0 −1 2 0 0 0 ]
[ ]
([ 0 0.5 0.5 0.2𝑎 −0.5 + 0.2𝑏 −1 + 0.2𝑐])

1 + 0.1𝑎 0.1𝑏 0.1𝑐

𝑛 + rank (𝐸𝐾 ) = 9 + rank ([ ]) .
0.2𝑎 −0.5 + 0.2𝑏 −1 + 0.2𝑐

When 𝑎, 𝑏, and 𝑐 satisfy for example, 𝑎 = 0, 𝑏 = 1, 𝑐 = 0, we have

𝐸𝐾 0 0
rank ([ ]) = 𝑛 + rank (𝐸𝐾 ) = 11. (73)
1 + 0.1𝑎 0.1𝑏 0.1𝑐 𝐴 𝐸𝐾 𝐵2
rank ([ ]) = 2, (72)
[ 0.2𝑎 −0.5 + 0.2𝑏 1 + 0.2𝑐] Then closed-loop system (6) is impulse controllable.
Mathematical Problems in Engineering
In the practical example, we can find that our method
of derivative feedback can change impulse controllability
when the method of proportional feedback is unable to do
it. So, derivative feedback method is more effective than
proportional feedback in the problem of impulse elimination.
5. Conclusions
In this paper, the problem of impulse controllability and
impulse observability of large-scale interconnected descrip-
tor system by derivative feedback has been studied. Necessary
conditions for the existence of a derivative feedback con-
troller for the first subsystem of the large-scale interconnected
descriptor systems ensuring the second subsystem to be
impulse controllable and impulse observable are derived,
respectively. The examples have been presented to demon-
strate the applicability of the proposed approach. Recently,
to ensure the reliability and safety of modern large-scale
industrial processes, data-driven methods have been receiv-
ing considerably increasing attention. The research on the
modeling and the impulsive controllability/observability of
interconnected descriptor systems based on large data and
fault diagnosis of complex system will be the next research
focus [31–33].
Conflict of Interests
The authors declare that there is no conflict of interests
regarding the publication of this paper.
Acknowledgments
This work is supported by National Natural Science Foun-
dation of China under Grant 61203001 and Fundamen-
tal Research Funds for Central Universities under Grant
N130405013.
References
[1] Q. L. Zhang, C. Liu, and X. Zhang, Complexity, Analysis and
Control of Singular Biological Systems, vol. 421 of Lecture Notes
in Control and Information Sciences, Springer, Berlin, Germany,
2012.
[2] D. M. Yang, Q. L. Zhang, and B. Yao, Descriptor Systems, Science
Press, Beijing, China, 2003.
[3] Z. G. Wu, H. Y. Su, and J. Chu, “Delay-dependent robust expo-
nential stability of uncertain singular systems with time delays,”
International Journal of Innovative Computing, Information and
Control, vol. 6, no. 5, pp. 2275–2283, 2010.
[4] Y.-F. Feng, X.-L. Zhu, and Q.-L. Zhang, “An improved 𝐻∞
stabilization condition for singular time-delay systems,” Inter-
national Journal of Innovative Computing, Information and
Control, vol. 6, no. 5, pp. 2025–2034, 2010.
[5] Y. F. Teng, Q. L. Zhang, and G. S. Zhang, “Study of controlla-
bility of descriptor systems,” Journal of Northeastern University,
supplement, pp. 1–6, 2003.
[6] D. Cobb, “Feedback and pole placement in descriptor variable
systems,” International Journal of Control, vol. 33, no. 6, pp. 1135–
1146, 1981.
13
[7] M. A. Shayman, “Homogeneous indices, feedback invariants
and control structure theorem for generalized linear systems,”
SIAM Journal on Control and Optimization, vol. 26, no. 2, pp.
387–400, 1988.
[8] J. J. Loiseau, “Pole placement and connected problems,” Kyber-
netika, vol. 28, no. 2, pp. 90–99, 1992.
[9] P. Zagalak, V. Kučera, and J. J. Loiseau, “Dynamics assignment
by PD state feedback in linear reachable systems,” Kybernetika,
vol. 30, no. 2, pp. 153–158, 1994.
[10] J. J. Loiseau and P. Zagalak, “On pole structure assignment in
linear systems,” International Journal of Control, vol. 82, no. 7,
pp. 1179–1192, 2009.
[11] J. J. Loiseaua, K. Özçaldiran, M. Malabre, and N. Karcanias,
“Feedback can onical forms of singular systems,” Kybernetika,
vol. 27, no. 4, pp. 289–305, 1991.
[12] J. Y. Ishihara and M. H. Terra, “Impulse controllability and
observability of rectangular descriptor systems,” IEEE Transac-
tions on Automatic Control, vol. 46, no. 6, pp. 991–994, 2001.
[13] M. Hou, “Controllability and elimination of impulsive modes
in descriptor systems,” IEEE Transactions on Automatic Control,
vol. 49, no. 10, pp. 1723–1727, 2004.
[14] G. R. Duan and A. G. Wu, “I-controllablizability in descriptor
linear systems,” Dynamics of Continuous, Discrete and Impulsive
Systems, vol. 13, no. 1, pp. 1197–1204, 2006.
[15] G.-R. Duan, A.-G. Wu, Y. Zhao, and H.-H. Yu, “A revisit to I-
controllablisability for descriptor linear systems,” IET Control
Theory & Applications, vol. 1, no. 4, pp. 975–978, 2007.
[16] A.-G. Wu, G.-R. Duan, and S.-M. Zhao, “Impulsive-mode
controllablisability in descriptor linear systems,” IET Control
Theory & Applications, vol. 1, no. 3, pp. 558–563, 2007.
[17] A.-G. Wu, G.-R. Duan, and H.-H. Yu, “Impulsive-mode con-
trollablizability revisited for descriptor linear systems,” Asian
Journal of Control, vol. 11, no. 3, pp. 358–365, 2009.
[18] D. L. Chu and D. W. C. Ho, “Necessary and sufficient conditions
for the output feedback regularization of descriptor systems,”
IEEE Transactions on Automatic Control, vol. 44, no. 2, pp. 405–
412, 1999.
[19] D. H. Wang and C. B. Soh, “On regularizing singular systems
by decentralized output feedback,” IEEE Transactions on Auto-
matic Control, vol. 44, no. 1, pp. 148–152, 1999.
[20] G. S. Zhang and W. Q. Liu, “Impulsive mode elimination
for descriptor systems by a structured P-D feedback,” IEEE
Transactions on Automatic Control, vol. 56, no. 12, pp. 2968–
2973, 2011.
[21] M. Hou and P. C. Muller, “Causal observability of descriptor
systems,” IEEE Transactions on Automatic Control, vol. 44, no.
1, pp. 158–163, 1999.
[22] Q. L. Zhang, Singular Large-Scale Systems Decentralized Control
and Robust Control, Northwest Industry University Press, Xian,
China, 1997.
[23] D. D. Siljak, Large-Scale Dynamic Systems: Stability and Struc-
ture, North-Holland, Amsterdam, The Netherlands, 1978.
[24] P. Lu, Y. Yang, Z. Li, and L. Huang, “Decentralized dynamic
output feedback for globally asymptotic stabilization of a class
of dynamic networks,” International Journal of Control, vol. 81,
no. 7, pp. 1054–1061, 2008.
[25] W. J. Xiong, D. W. C. Ho, and J. D. Cao, “Synchronization
analysis of singular hybrid coupled networks,” Physics Letters A,
vol. 372, no. 44, pp. 6633–6637, 2008.
[26] Y. J. Zhou, X. H. Yu, C. Y. Sun, and W. W. Yu, “Robust synchroni-
sation of second-order multi-agent system via pinning control,”
IET Control Theory & Applications, vol. 9, no. 5, article 775, 2015.
14 Mathematical Problems in Engineering

[27] S. Keshmiri, R. Colgren, and M. Mirmirani, “Six-DOF of

freedom nonlinear equations of motion for a generic hyper-
sonic vehicle,” in Proceedings of the AIAA Atmospheric Flight
Mechanics Conference and Exhibit, Hilton Head Island, SC,
USA, August 2007.
[28] G. Y. Peng, L. J. Peng, S. W. Lin, and F. A. Zhang, Anti-
Aircraft Missile Guidance Control System, China Astronautic
Press, Beijing, China, 2009.
[29] F.-Q. Zhou, Y. Wang, J. Zhou, and J.-G. Guo, “Design of variable
structure controller for hypersonic coupled vehicle system,”
Journal of Astronautics, vol. 32, no. 1, pp. 66–71, 2011.
[30] M. D. S. Aliyu and E. K. Boukas, “Kalman filtering for affine
nonlinear descriptor systems,” Circuits, Systems, and Signal
Processing, vol. 30, no. 1, pp. 125–142, 2011.
[31] S. Yin, S. X. Ding, X. Xie, and H. Luo, “A review on basic data-
driven approaches for industrial process monitoring,” IEEE
Transactions on Industrial Electronics, vol. 61, no. 11, pp. 6414–
6428, 2014.
[32] S. Yin, X. Zhu, and O. Kaynak, “Improved PLS focused on key-
performance-indicator-related fault diagnosis,” IEEE Transac-
tions on Industrial Electronics, vol. 62, no. 3, pp. 1651–1658, 2015.
[33] S. Yin and Z. H. Huang, “Performance monitoring for vehicle
suspension system via fuzzy positivistic c-means clustering
based on accelerometer measurements,” IEEE/ASME Transac-
tions on Mechatronics, vol. 20, no. 5, pp. 2613–2620, 2015.
Advances in Advances in Journal of Journal of
Operations Research
Hindawi Publishing Corporation
Decision Sciences
Hindawi Publishing Corporation
Applied Mathematics
Hindawi Publishing Corporation
Algebra
Hindawi Publishing Corporation
Probability and Statistics
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

The Scientific International Journal of

World Journal
Hindawi Publishing Corporation
Differential Equations
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

Submit your manuscripts at

http://www.hindawi.com

International Journal of Advances in

Combinatorics
Hindawi Publishing Corporation
Mathematical Physics
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

Journal of Journal of Mathematical Problems Abstract and Discrete Dynamics in

Complex Analysis
Hindawi Publishing Corporation
Mathematics
Hindawi Publishing Corporation
in Engineering
Hindawi Publishing Corporation
Applied Analysis
Hindawi Publishing Corporation
Nature and Society
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

International
Journal of Journal of
Mathematics and
Mathematical
Discrete Mathematics
Sciences

Journal of International Journal of Journal of

Hindawi Publishing Corporation Hindawi Publishing Corporation Volume 2014

http://www.hindawi.com
Function Spaces
Hindawi Publishing Corporation
Volume 2014 http://www.hindawi.com http://www.hindawi.com
Stochastic Analysis
Hindawi Publishing Corporation
Optimization
Hindawi Publishing Corporation
Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014