ISSN 2347-1921
REGIONAL GRADIENT STRATEGIC SENSORS
CHARACTERIZATIONS
Raheam A. Al-Saphory1*, Naseif J. Al-Jawari2 and Assma N. Al-Janabi2
1
Department of Mathematics, College of Education for Pure Sciences, Tikrit University, Tikrit, IRAQ.
1
Email: saphory@hotmail.com
2
Department of Mathematics, College of Science, Al-Mustansiriyah University, Baghdad, IRAQ.
2
njaljawari@hotmail.com
asm_no_2006@yahoo.com
2
ABSTRACT
In the present paper, the characterizations of regional gradient strategic sensors notions have been given for different
cases of regional gradient observability. The results obtained are applied to two dimensional linear infinite distributed
system in Hilbert space where the dynamic is governed by strongly continuous semi-group. Various cases of regional
strategic sensors are considered and analyzed in connection with regional gradient strategic sensors concepts. Also, we
show that there is a various sensors which are not gradient strategic in usual sense for the considered systems, but may
be regionally gradient strategic of this system.
Keywords:
-strategic sensors; exactly
-observability; weakly
-observability; diffusion systems.
Academic Discipline And Sub-Disciplines:
Applied Mathematics/ Control Systems and Analysis
2010 AMS SUBJECT CLASSIFICATION:
93A30; 93B07; 93B28; 93C05; 93C28.
TYPE (METHOD/APPROACH)
Mathematical approach for parabolic distributed parameter systems gvorned by semi-group operator in a Hilbert state
space.
Council for Innovative Research
Peer Review Research Publishing System
Journal: JOURNAL OF ADVANCES IN MATHEMATICS
Vol . 10, No. 6
www.cirjam.com , editorjam@gmail.com
3605 | P a g e
May 25, 2015
ISSN 2347-1921
1. NTRODUCTION
The analysis of distributed parameter systems refers to a set of concepts such as controllability, observability, detectability
[13-14, 18]. The study of these concepts can be made via actuators and sensors structures see [14-17], these concepts
give an important link between a system and it’s environment [15-18], so that the concepts of actuators and strategic
sensors for a class of distributed parameter systems are introduced in order that controllability and observability can be
achieved [14-18]. The regional analysis is one of the most important notion of system theory [20-22], it consist to
reconstruction the state observation on a sub-region
of spatial domain Ω in finite time [19-23, 25-28], this concepts
introduced and developed by El-Jai et al. An important extended to the asymptotic case for infinite time by El-Jai and AlSaphory in several works [1-7]. The study of regional gradient observability for a diffusion system has been given in [2728] where one is interested in knowledge of the state gradient only in a critical sub-region of the system domain without
the knowledge of the state itself. Moreover, the applications are motivated by many real world see [10-12, 22]. Commercial
buildings are responsible for a significant fraction of the energy consumption and greenhouse gas emissions in the U.S.
and worldwide. Consequently, the design, optimization and control of energy efficient buildings can have a tremendous
impact on energy cost and greenhouse gas emission. Mathematically, building models are complex, multi-scale, multiphysics, highly uncertain dynamical systems with wide varieties of disturbances [10].
Fig. 1: Room control model with sensor, in flow and out flow
In this paper we use a model problem to illustrate that distributed parameter control based on PDEs, combined with high
performance computing can be used to provide practical insight into important issues such as optimal sensor/actuator
placement (may be best or strategic sensors/ actuators) and optimal supervisory building control. In order to illustrate
some of the ideas, we consider the problem illustrated by a single room shown in (Figure 1). This model one can
reformulated [11] as spatial case of more general model of distributed parameter systems and represented in the next
section (see Figure 2). In addition, the characterization of regional strategic sensors have been given for various types of
regional observability in [7].
The purpose of this paper is to extended these results in [7] to the case of regional gradient sensors. Thus, we give a
characterization of regional gradient strategic sensors for different cases of regional gradient observation. Therefore, we
study and analyze the relationship between the regional gradient strategic sensors and the regional exactly gradient
observability. So, the outline of this paper is organized as follows:
Section 2 is present problem statement and basic definitions with characterization of the regional gradient observability.
The mathematical concepts of regional gradient strategic sensors in a various situations are studied and developed in
section 3. In the last section we gives an application about different sensors locations.
2. REIONAL GRADIENT OBSERVABILITY
In this section, we are interested to recall the notion of regional gradient observability and give original results related to
particular systems as in [27-28].
2.1 Problem Statement
Let Ω be a regular bounded open subset of
, with a smooth boundary Ω and be a non-empty given sub-region of Ω.
Let 0, , > 0 be a time of measurement interval. We denoted = Ω × 0,
and � = Ω ×]0, T[. Consider the following
distributed parabolic defined by
�,
�,
=
�, 0 =
�, = 0
+
in
�
0
in Ω
in �
(1)
with the measurements given by the output function
We have
∙,
=
=
,
, =1
(
3606 | P a g e
(2)
), with
�( ).
May 25, 2015
ISSN 2347-1921
Suppose that - is elliptic, i.e., there exists
��
, =1
=1
�
2
> 0 such that
, almost everywhere (a.e) on , ∀� = �1 , … , �
.
This operator is a second order linear differential operator, which generator a strongly continuous semi-group
1
0
on
the Hilbert space = (Ω) and is self-adjoint with compact resolvent. The operator
,
and
,
,
depend on the structure of actuators and sensors [18]. The space , and � be separable Hilbert spaces where is the
state space, = 2 0, ,
is the control space and � = 2 0, ,
is the observation space where and are the
numbers of actuators and sensors (see Figure 2).
Fig. 2: The domain of �, the sub-region �, various sensors locations
Under the given assumption, the system (1) has a unique solution [24]:
�,
=
0
� +
( − �)
0
�
�
(3)
The problem is to provide sufficient conditions to ensure that, how to extend the results in [7], so that to give a sufficient
conditions of regional gradient strategic sensors which achieve the observability in sub-region
using various regional
gradient sensors.
2.2 Definitions And Characterizations
The regional gradient observability concept and reconstruction gradient state have been introduced by Zerrik E. et al. as in
ref.s [27-28] and recently this concept is developed to the regional asymptotic case by Al-Saphory R [1-7]. Consider the
autonomous system to (1) given by
�,
�,
=
in
�, 0 = 0 �
�, = 0
(4)
in Ω
in �
The solution of (4) is given by the following form,
�,
=
0
� ,
∀
(5)
0,
The measurements are obtained through the output function
.,
�,
=
We first recall a sensors is defined by any couple (D, f ), where D is spatial support represented by a nonempty part of Ω
and f represents the distribution of the sensing measurements on D.
Depending on the nature of D and f, we could have various type of sensors. A sensor may be pointwise if D = {b} with
Ω and = (. − ), where is the Dirac mass concentrated at b. In this case the operator C is unbounded and the
output function (2) can be written in the form [13-14]
= ( , )
It may be zonal when
�,
=
⊂ Ω and
2(
). The output function (2) can be written in the form
�
Now, we define the operator
→
:
=
Thus, we get that
∙,
=
3607 | P a g e
.
�
(6)
(∙ ,0)
May 25, 2015
ISSN 2347-1921
where K is bounded linear operator [8].
:� →
We note that
=
is the adjoint operator of K defined by
(7)
0
Consider the operator
1
:
→
It's adjoint
→ ( 2 ( Ω))
=(
�1
,…,
is given by
�
(8)
)
Ω)) → 1 (Ω)
→
=
is a solution of the Dirichlet problem
:
where
Ω
(
2(
∆ = −div
=0
in Ω
in Ω
For a nonempty subset
:
and
(9)
of Ω, we consider the operators
( 2 ( Ω)) → ( 2 ( ))
→
= ∣
( Ω) → 2 ( )
→
= ∣
∣ is the restriction of
(10)
2
(11)
:
where
[9].
⋄ Their adjoints are respectively denoted by
( 2 ( ))
:
→
and
=
)→
→
and
are defined by
( 2 ( Ω))
∣ in
0 in Ω
(12)
Ω)
(13)
The
∣
in
→
=
0 in Ω
idea of gradient observability is based on the existence of an operator : � → ( 2 ( )) such that
= 0 . This is a
natural extension of the observability concept [8]. Then we defined the operator =
from � into ( 2 ( )) as in
[27]. Now, let us denoted the system (4) together with the output (2) by (4)-(2).
2(
:
2(
Definition 2.1: The system (4)-(2) is said to be regionally exactly observable on a sub-region
(exactly
-
observable), if
=
2
( )
Definition 2.2: The system (4)-(2) is said to be regionally weakly observable on
(. ) =
2(
(weakly
- observable), if
)
Definition 2.3: The system (4)-(2) is said to be regionally exactly gradient observable on
(exactly
- observable), if
= ( 2 ( ))
Definition 2.4: The system (4)-(2) is said to be regionally weakly gradient observable on
(weakly
- observable),
if
(. ) = ( 2 ( ))
We see that if a system is weakly
gradient, viz., if
– observable then there is one to one relationship between the output and the initial
is given and 0 satisfies
=
.
0 , then
0 is a unique.
Remark 2.5: We can deduced that, the definition 2.4 is equivalent to say that the system (4)-(2) is weakly
observable, if
3608 | P a g e
May 25, 2015
-
ISSN 2347-1921
= {0}
Then, the following characterization can extend to the regional gradient case as in ref. [26].
Proposition 2.6: The system (4)-(2) is exactly
-observable if and only if there exist
> 0 such that for all
( 2 ( )) , such that,
(14)
�
( 2 ( ))
Proof: The proof of this property is deduced from the usual results on observability considering
,
be a reflexive Banach space and let
and
,
,
,
⊂
.
There exist > 0 such that
1.
2.
[13]. Let
, then the following conditions are equivalent
∀
Now, by applying the above result we obtain the equivalent condition for exactly
Let
= �,
= ( 2 ( )) ,
=
=
Now, since the system is exactly
such that
2(
2(
(
and
))
.
=
-observable we have
∀
))
Remark 2.7: We have:
-observable as:
.□
⊂
, which is equivalent to that fact there exist
> 0,
(1) The regional state reconstruction will be more precise than the whole domain if we estimate the state in the whole the
domain.
(2) From (14) there exists a reconstruction error operator that gives the estimation
If we put e = 0 − 0 , we have
e
e
( 2 ( ))
⟹
Where
−
0
0
0 (
2(
0
))
−
0 (
Proof: Since the system is exactly
since (
2
0 ( 2 (Ω))
2
={
=
0 (
0:
0
2
To prove
> 0 and
0
=
2(
2(
0
))
< ∞},
0 ( 2 ( ))
, and then,
-observable then it is exactly
< ∞,
0
=
2
0, ,�
), ∀
0
�
0
2
> 0 such that ∀
-observable.
0
2
( ), we have
( ), where
(15)
∀ = 1,2, … .
then from (15) and since a system exactly
-observable, then there exist
1
> 0 such that
=
in
is the estimated state of the system.
, ∀ >0
0, ,�
2
:
0
-observable there exist
, then
and then
( 2 ( )) =
2
0
)
)) ⊂
0
2 (Ω))
is the exact state of the system and
2(
of the initial state
( 2 (Ω))
Proposition 2.8: If the system is exactly
0
2(
0
= . By choosing
0 �
(16)
0 �
Then, we can get
0 ( 2 ( ))
0
(17)
0 �
2
And by substituting (16) in (17), we obtain
0 ( 2 ( ))
0 �
Therefore this system is exactly
3609 | P a g e
-observable with
= 1.
May 25, 2015
ISSN 2347-1921
Remark 2.9: From the above proposition we can get the following result:
If the system is exactly
1
-observable then it is exactly G-observable in
1
for all
3. REGIONAL GRADIENT STRATEGIC SENSORS
⊂
(exactly
1
-observable).
The purpose of this section is to give the characterization for sensors in order that the system (4)-(2) which is observable
in .
3.1 �� -Strategic Sensors
Definition 3.1: A sensor ( , ) is regional gradient strategic on
(
-strategic) if the observed system is weakly
-
observable.
Definition 3.2: A sensor ( , )1
is
-strategic if there exist at least one sensor (
1, 1 )
which
-strategic.
We can deduce that the following result:
Corollary 3.3: A sensor is
-strategic if the observed system is exactly
Proof: Let the system exactly
-observable.
-observable, then , we have
= ( 2 ( ))
From decomposition sub-space of direct sum in Hilbert space, we represent ( 2 ( Ω)) by the unique form [13]
= ( 2 ( Ω))
+
We obtain
= {0}
This is equivalent to [9]
(. ) = ( 2 ( ))
Finally, we can deduce this system is weakly
Corollary 3.4: A sensor is
Proof: Since a sensor is
- observable and therefore this sensor is
-strategic if and only if the operator � =
-strategic.□
is positive definite.
-strategic this mean that the system is weakly
- observable,
( 2 ( )) such that
let
<�
,
>(
2(
))
and therefore
= 0 then
( 2 ( )) such that
= 0, then <
,
>� = 0
and thus,
Hence
,
=0
= 0, i.e., � is positive definite.
, thus,
Conversely, let
<�
�
>(
2(
))
=0
= 0 thus the system is weakly
- observable and therefore a sensor is
-strategic.□
Remark 3.5: From the previous results, we obtain that:
(1) If the system is exactly
strategic.
-observable then the system is weakly
(2) A sensor which is regional gradient strategic sensor in
gradient strategic sensor in 2 ( 2 -strategic) for any 2 ⊂ 1 .
(3) The concept of exact
1
(
1
-observable and therefore this sensor is
-strategic) for a system where
-observability is more restrictive than weak
1
-
⊂ Ω, is regional
-observability.
Now, assume that the operator has a complete set of eigenfunction in 1 Ω , denoted by
, which is orthonormal in
2
( ) and the associated with the eigenvalue � of multiplicities , then the concept of regional gradient strategic on
can be characterized by the following result:
Theorem 3.6: Assume that sup
3610 | P a g e
=
< ∞, then the suite of sensors ( , )1
,
-strategic if
May 25, 2015
ISSN 2347-1921
(1)
(2)
∀
=
Where
=(
(
)
1,
for 1
=1
) =
=1
, and
,1
in the pointwise case
( )
�
<
>
,
�
in the zonal case
2
Proof: We will discussed the case where the sensors are of pointwise type and located inside the domain Ω. The suite
of sensors ( , )1
is
-strategicif and only if
( 2 ( ))
<
,
>(
2(
= 0, ∀
))
Suppose that the suite of sensors ( ,
=(
<
1
0,
2
>
<
Let
,
0
2(
,…,
)=
0,
=
=1
<
,
>(
0
)=
≠
,
0
))
=
0
2(
2(
x
-strategicbut for a certain
�,
, then there exists a vector
0. So, we can construct a nonzero
0
2(
) considering
.
,1
= ( 0,
=1
<
0 , … , 0 ),
�
=1
<
�
then
,
0
(
>
2(
),
)
0
>
2
Ω
is the solution of the following system:
�,
�,
=
�, 0 = 0
�, = 0
+
−
=1
in
�,
=−
�, 0 =
�, = 0
Multiply (18) by
�
�,
=1
�
0
�,
(18)
in Ω
in �
Consider the system:
Q
= 0.
and
=
where
is
)1
≠ 0, such that
)
>
� ⟹
in
(19)
in Ω
in �
and integrate on Q, we obtain
�,
�
( − )
�
�,
�,
�
�,
�
�
+
.
But we have
�
�,
�,
�
Ω
�
�,
then:
Ω
�
�,
�,
�
integrating by parts we obtain
3611 | P a g e
�
+
Ω
�,
�,
�
�,
=1
� +
0
�,
�,
�+
�+
( − )
�,
�
�
�,
�,
�,
�
�
�
�,
℥
�,
�,
�
�
May 25, 2015
Ω
�
�
�
�,
�,
� −
�,
ƞ
ƞ,
the boundary conditions give
Thus
Ω
�,
�
�,
Ω
�
�,
�,
�
and we have
<
,
0
2(
(ƞ, )
ƞ
�
))
=
�,
Then
−� ( − )
∞
=1
=
=
,
=1 �
<
=1
∞
=1
−� ( − )
=
∞
=1
�
�,
>
=1
−
<
(
�,
�
.
2(
�
( − ) .
,
=1 �
=1 0
°,
�
�
( − )
�,
�
=−
But
( − )
�,
�
,
�
=1 Ω
ƞ
=1
( − )
=1 0
ƞ,
�
+
=1
�=−
>(
ISSN 2347-1921
,
)
0,
>
2(
)
=1 �
)
therefore
<
,
0
>(
2(
))
Thus
<
,
0
This is true for all
strategic.□
>(
2(
2
))
=−
=−
(0, ;
=1 0
=1 0
), then
∞
=1
�
0
−
�
−
(
(
)
)
( − )
−
(20)
=0
which contradicts the assumption that the suite of sensors is
-
We can deduced the following result:
Corollary 3.7: In the one dimension case, a sensor is
,∀
1, where
-strategic if and only if
=
and
=
is given in theorem 3.6.
Remark 3.8: From the previous results, we can get
(1) The Theorem 3.6 implies that the required number of sensors is greater than or equal to the largest multiplicity of
the eigenvalues.
(2) By infinitesimally deforming the domain, the multiplicity can be reduced to one [19]. Consequently,
-strategic
sensors can be achieved using only one sensor.
Now, we can deduced that various sensors which are not strategic in usual sense for systems, but may be
strategic and achieve the
-observability. This is illustrated in the following counter- example.
-
3.2 A Counter- Example
Consider the system described by the parabolic equation
�,
2
=
0,1 =
� ,0 =
�2
�,
1,
0 �
in ]0,1 × 0, [
=0
in ]0, [
in ]0,1[
Suppose that the measurement is given by pointwise sensor located in
function
3612 | P a g e
(21)
]0,1[ which is given by the following output
May 25, 2015
ISSN 2347-1921
.,
=
(22)
�,
Ω
�−
�=
,
,
(0, )
Where
= 2 sin �� and � = − 2 � 2 . First, we must prove that the system (21)-(22) is not weakly observable in Ω,
that means the sensors ( , ) is not strategic. For this purpose, we can write the system (21) as a state space one
dimensional system
�,
Where
�,
=
� ,0 =
2
=
�2
0
�
0
generate the continuous semigroup (
∞
=1
=
�
<
0,
>
∞
=1
=
�
<
0,
>
2�2
given by [17].
2
are the eigenfunctions associated with the eigenvalues of
=
Ω
= ( )
0
The system (21)-(22) is weakly observable if
. Then from solution
0
= {0}.
then system (21)-(22) is not weakly observable on Ω=(0,1) and a sensor ( , ) is not strategic.
As proved in [27], if
A sensor is
0
2 (Ω )
Where,
= 2 sin �� , � = −
of (21), we have
�,
)
⇔ :
-strategic on (0,1)
∞
=1
=
[1, − 1] ∩ � . Since sin �
=0⇔
=
condition
in
⟹
= .
2 +1
∞
=1 2
Consequently, the system is weakly observable on (0,1). And then, it is G-strategic on (0,1) ⇔
=
[0, − ∩�. Since cos � = ⇔ =2 +12⟹ =2 +12 . Consequently, the system is weakly G-observable on (0,1).□
Corollary
3.9: If the
(3.6) is satisfied and a sensor is
system (21)-(22)
-strategic.
is
exactly
-observable,
rank
Now, assume that a sensor is not gradient strategic in whole the domain Ω and let ( )
⊂ � be such that
=
{( ) } and = � \ .
�
be a basis in (
Proposition 3.10: The following properties are equivalent:
1.
2.
3.
A sensor is
-strategic.
)
{(
) }=( 2
2
If
(
) is such that < ,
4.
If
= 0 in Ω\ , then
Proof: 1⟹2 Since sensors are
Let
2
(
−
=
=
−
<
�
(
2
>(
2
we deduced that
)
3⟹4 Let
3613 | P a g e
<
)
2
Ω )
)
=
, then
Ω ) . Let
= 0.
.
< ,
. Then
2
)
>0∃
(
(
) . For any
(
2
>�
>�
.
}
2
−
2
>(
< ,
{
2⟹3 Let
2
(
,
= 0 for all
)
2
-observable.
� such that
, but
)
< ,
and hence
< ,
2
(
2
-strategic this mean that the system is weakly
>0∃
) . Then for
>(
= 0 for all
theorem
>�
, and thus
<
) such that
< , with
= 0, ∀
. Thus,
= 0.
= 0 in Ω\ .
May 25, 2015
ISSN 2347-1921
Now Consider
< ,
Since
>(
=
2
)
). For
(
=
<
,
>(
, we have
= 0, we get
= 0 in Ω and
4⟹1 Consider
2
(
= 0, ∀
2
(
then
< ,
>(
I.
) such that
= 0. We have
=
= 0.
Ω )
2
�
< ,
2
Ω )
>(
2
=
)
{( )
)
(
< ,
>(
. Hence
= 0.□
2
) = 0. Therefore,
)
}
and then
< ,
>(
2
= 0, ∀
)
Therefore
=
< ,
>(
2
.
= 0 in Ω\ .
)
From the assumption we have < ,
>(
2
= 0, ∀
)
We can deduced the following result:
Corollary
3.11:
⊂ Ω such that <
Under the hypotheses of
, ≠ .
>( 2 ) = 0, ∀ ,
,
Proposition
3.10,
a
= 0, ∀
< ,
Since
0
. Let
=
Ω )
=<
>(
2
and
in
,
0
>(
2
I. Then
0
Ω )
=
2
)
<
,
0
>(
= 0 in Ω\ , under the assumption of Corollary 3.11 we have
=
<
From (23)-(24), we obtain
= 0 ,∀
< ,
0
>(
2
Ω )
,
0
>(
=
2
0 (
0
2
2
Ω )
is
-strategic
in
all
= 0 in Ω\ . Then we only need to show
Proof: To deduce the result from previous Proposition 3.10, we take
that
sensors
=
0
)
(23)
(24)
.
4. APPLICATION TO SENSORS LOCATIONS
In this section, we give specific results related to the different case presented in the above section. First we consider
internal sensors (zonal, pointwise, filament in rectangular and disk domain) the presented result give information on the
structure of . Consider the system
�1 , �2 ,
= ∆ �1 , �2 ,
in
,
in Ω,
�1 , �2 , 0 = 0 �1 , �2
�1 , �2 , = 0
in �
Let Ω = (0,1) × (0,1) and let = ( 1 , 1 ) × ( 2 ,
eigenvalue of the system (25) are given by:
2
�1 , �2 =
1− 1
Associated with eigenvalue
� =−
sin �
(�1 −
1)
1− 1
2)
�
be the considered region is subset of Ω, the eigenfunctions and the
(�2 −
2)
2− 2
2
2
1−
2− 2
(25)
1
2
+
2−
2
(26)
(27)
2
4.1 Internal Zone Sensor
Consider the system (25) together with output function (2) where the sensor supports
can by written by the form
=
3614 | P a g e
�1 , �2 ,
�1 , �2
�1 �2
are located in Ω. The output (2)
(28)
May 25, 2015
ISSN 2347-1921
Where ⊂ Ω is location of zone sensor and
eigenvalues (26) and (27).
2(
). In this case of (see Figure 3), the eigenfunctions and the
Fig. 3: Domain Ω, sub-region � and location � of internal zone sensor
However , if we suppose that
1− 1
2− 2
2
2
Then multiplicity of � is
= 1 and then one sensor ( , ) my be sufficient to achieve
and (28) [19]. Let the measurement support is rectangular with
= [�01 − 1 , �01 + 1 ] × [�02 − 2 , �02 + 2 ]
Ω
Then, we have the following result
Corollary 4.1:
If
1
strategic if
(�01 −
1)
1− 1
and
(�02 −
1)
2− 2
-observable of the systems (25)
is symmetric about �1 = �01 and
2
is symmetric about �2 = �02 , then the sensor
,
is
-
� for some , .
4.2 Internal Pointwise Sensor
In this case the out put function is given by:
=
With
If
=(
1
,
1− 1
2− 2
2)
�1 , �2 ,
�1 −
1− 1)
1− 1
, then
and
−
2
�1 �2
(29)
is location of pointwise sensor as defined in (see Figure 4)
Fig. 4: Rectangular domain, and location � of internal pointwise sensor
= 1 and one sensor ( ,
Corollary 4.2: The sensor
(
1 , �2
(
2− 2)
2− 2
,
is
) may be sufficient for
-observability of the systems (25)-(29)
-strategic if
�, for some , .
4.3 Internal Filament Sensor
Consider the case where the observation is given on the curve � =
3615 | P a g e
with
1
(0,1) (see Figure 5)
May 25, 2015
ISSN 2347-1921
Fig. 5: Rectangular domain, and location � of internal filament sensors
Corollary 4.3: If the measurements recovered by filament sensor (�, � ) such that is symmetric with respect to the line
� = �0 . Then the sensor (�,
(�01 −
1)
1− 1
and
(�02 −
2)
2− 2
�)
is
-strategic if
�, for , = 1, … , .
Remark 4.4: These results can be extended to the following:
1. Case of Neumann or mixed boundary conditions [4-5].
2. Case of disc domain
( D,1)
3. Case of boundary sensors where
and
(0, r )
where
and
0 r 1
C L( X , R q ) , we refer to see [13-14].
[1-3] .
4. We can show that the observation error decreases when the number and support of sensors increases [23, 25].
5. CONCLUSION
We have been introduced a sufficient condition of regional gradient strategic sensors in order to achieves regional gradient
observability. Many interesting results concerning the choice of sensors structure are given and illustrated in specific
situations. Various questions still opened under consideration. For example, these result can be extended to the boundary
case with parabolic and hyperbolic systems [8].
ACKNOWLEDGMENTS
Our thanks in advance to the editors and experts for considering this paper to publish in this estimated journal and for
their efforts.
REFERENCES
[1] Al-Saphory, R. and El Jai, A. 2001. Sensor structures and regional detectability of parabolic distributes systems.
International Journal of Sensors and Actuators. Vol. 90.163-171.
[2] Al-Saphory, R. and El Jai, A. 2001. Sensors characterizations for regional boundary detectability in distributed
parameter systems. International Journal of Sensors and Actuators. Vol. 94.1-10.
[3] Al-Saphory, R. and El Jai A. 2002. Regional asymptotic state reconstruction. International Journal of Systems Science.
Vol. 33. 1025-1037.
[4] Al-Saphory, R. 2010. Strategic sensors and regional exponential observation in Neumann boundary conditions. Zanco
Journal of Pure and Applied Science. Vol. 22, 1-11.
[5] Al-Saphory, R., Al-Jawari, N. and Al-Qaisi, A . 2010. Regional gradient detectability for infinite dimensional
systems. Tikrit Journal of Pure Science. Vol. 15, No. 2. 1-6.
[6] Al-Saphory, R. 2011. Strategic sensors and regional exponential observability. ISRN Applied Mathematics. Vol. 2011.
Article ID 673052. pp. 1-13.
[7] Al-Saphory, R., Al-Joubory, M. and Jasim, M. 2013. Regional strategic sensors characterizations. Journal of
Mathematical and Computational Science. 3, 401-418.
[8] Boutoulout, A., Bourray, H. and Khazari, A. 2014. Flux observability for hyperbolic systems. Applied Mathematics and
Information Sciences Letters. Vol. 2 , No. 1, 13-24.
3616 | P a g e
May 25, 2015
ISSN 2347-1921
[9] Brezis, H. 1987. Analyse fonctionnalle; Theorie et applications. 2em tirage. Masson. Paris, France.
[10] Burns, J., Jorggaard, J., Cliff, M. and Zietsman, L. 2009. A PDE approach to optimization and control of high
performance buildings, Workshop in Numerical Techniques for Optimization Problems with PDE Constraints, No. 04/2009,
DOI: 10.4171/OWR/2009/04. 25-31 January. Houston, Denmark.
[11] Burns, J., Jorggaard, J., Cliff, M. and Zietsman, L. 2010. Optimal sensor design for estimation and optimization of
PDE Systems, 2010 American Control Conference. Marriott Waterfront, Baltimore, MD, USA. June 30-July 02, 2010.
[12] Chen Y., Moore K., and Song, Z. 2004. Diffusion boundary determination and zone control via, mobile actuatorssensors networks (MAS-net)-challenge and opportunities. SCOIS. Utah state university, Logan, UT84322, USA.
[13] Curtain, R. F. and Pritchard, A. J. 1978. Infinite dimensional linear theory systems; Lecture Note in Control and
Information Sciences. Spring – Verlag, New York.
[14] Curtain, R. F. and Zwart, H. 1995. An introduction to infinite dimensional linear system theory. Springer-Verlag, New
York.
[15] El Jai, A. and Pritchard, A. j. 1987. Sensors and actuators in distributed system. International Journal of Control. Vol.
46, 1139-1153.
[16] El Jai, A. and Amouroux, M. 1988. Sensors and observers in distributed parameter systems. International Journal of
Control. Vol. 47, 333-347.
[17] El Jai, A. 1991. Distributed systems analysis via sensors and actuators, International Journal of Sensors and
Actuators.Vol. 29, 1-11.
[18] El Jai, A.. and Pritchard A. j. 1988. Sensors and controls in analysis of distributed system, Ellis Harwood series in
Mathematics and Applications, Wiley, New York.
[19] El Jai, A. and El Yacoubi, S. 1993. On the number of actuators in a parabolic system, Applied Mathematics and
Computer Science. Vol.3, No.4, 139-150.
[20] El Jai, A., Simon, M.C. and Zerrik, E. 1993. Regional observability and sensor structures, International Journal of
Sensor and Actuator. Vol. 39, No. 2, 95-102.
[21] El Jai, A., Amouroux, M. and Zerrik, E. 1994. Regional observability of distributed systems. International Journal of
Systems Science. 1994, Vol. 25, No.2, 301-313.
[22] El Jai, A., Simon, M.C., Zerrik, E. and Amouroux, M. 1995. Regional observability of a thermal proces. IEEE
Transaction on Automatic Control. Vol. 40, 518-521.
[23] El Jai A. and Hamzauoi, H. 2009, Regional observation and sensors. International Journal of Applied Mathematics
and Computer Sciences. Vol. 19, No. 1, 5-14.
[24] Gressang, R. and Lamont, G.1975. Observers for systems characterized by semi-group. IEEE on Automatic and
Control. 20, 523-528.
[25] Hamzauoi, H., 2008, On choice of actuators-sensors structures in the systems, Ph. D. Thesis, Perpignan, France.
[26] Zerrik, E. 1993. Regional analysis of distributed parameter systems, Ph.D. Thesis. University of Rabat, Morocco.
[27] Zerrik, E., Bourray H. and Badraoui L. 2000. How to reconstruct a gradient for parabolic systems. Conference of
MTNS 2000. Perignan, France, June-2000, 19-23.
[28] Zerrik, E. and Bourray, H. 2003. Gradient observability for diffusion systems. International Journal of Applied
Mathematics and Computer Sciences. Vol. 2, 139-150.
3617 | P a g e
May 25, 2015
ISSN 2347-1921
Author’s Biographies:
Raheam Al-Saphory is professor associated at the TIKRIT University in
IRAQ. He received his Ph.D. degree in Control System and Analysis in
(2001) from LTS/ Perpignan University France. He has a post doctorate as a
researcher in 2001-2002 at LTS. Al-Saphory wrote one book and many
papers in the area of systems analysis and control. Also he is a supervisor of
many Ph D. and Msc. students and he was Ex-head of Departement of
Mathematics /College of Eduction for Pure Science Tikrit University 20102011. He visite many centres and Scientific Departments of Bangor
University/ Wales/ UK with academic staff of Iraqi Universities in 2013. He is
a coordinator of postgraduate studies at Department of Mathematics 2014present.
Naseif Al-Jawari is assistant Proffessor at the Al-Mustansiriyah
University, IRAQ. He received his Ph.D. degree in Optimal Control
Theory in (2000) from Faculty of Mathematics/ Łódź University
Poland. Al-Jawari wrote many papers in the area of systems analysis
and Optimization. Also he is a supervisor of many Ph D. and Msc.
students. He is head of Applied Mathematics Branch/ Department of
Mathematics/ College of Science/ Al-Mustansiriyah University 2012present.
Assma N Al-Janabi is an instructor and researcher at the Department of
Mathematics/ College Science/ Al-Mustansiriyah University/ IRAQ 2012Present. Here research area focused on Distributed Parameter Systems
Analysis and Control. She is obtained here Ms.c. degree from Al-Nahrain
University in (2003).
3618 | P a g e
May 25, 2015