Chen ch5

~
182
CONTROLLABILITY AND OBSERVABILITY OF LINEAR DYNAMICAL EQUATIONS

CON'
Definition 5-4
b(t)
The dynamical equation E is said to be uniformly controllable if and only if there

exist a positive u, and positive ai that depend on u, such that
O<al(u,)I~W(t,
and
t +u,);s;aiu,)I
0< a3(uJI ;s;cJ)(t +U,' t)W(t, t +u,)cJ)*(t +U,' t) ;s;a 4 (u,)I
-3
-2
-I
for all t in ( - 00, (0), where cJ) is the state transition matrix and W is as defined
in Equation (5-13).
Figure 5-5 The function b(t).
By A> B, we mean that the matrix (A - B) is a positive definite matrix (see

Section 8-5). Uniform controllability ensures that the transfer of the states
can be achieved in the time interval U c' The concept of uniform controllability
is needed in the stability study of optimal control systems. See References 56
and 102.
Instantaneous controllability and uniform controllability both imply con
trollability. However, instantaneous controllability neither implies nor is
implied by unifor.m controllability.
achieved in the time interval U c :

will not be arbitrarily large [see (
W - 1]. In optimal control thea
sometimes required to ensure th
Time-invariant case. In thi~
n-dimensional linear time-invaril
Example 4
FE:
Consider the one-dimensional linear dynamical equation
where x is the n x 1 state vector,

and n x p real constant matrices,
equations, the time interval of i
that is, [0, (0),
The condition for a time-vary
that there exists a finite t I such that
in t, on [to, t 1 ]. In the time-im
As discussed in Chapter 4, all eler
terms of form IkeA.t: hence they are
sequently, if its rows are linearly
independent on [to, t IJ for any to
time-invariant state equation is c(
and the transfer of any state to an~
time interval. Hence the referenc
trollability study of linear time-im
e-Itlu
Since p(Mo(t)) = p(e -Itl) = 1 for all t, the dynamical equation is instantaneously
controllable at every t. However, the dynamical equation is not uniformly
controllable because there exists no althat depends on u, bilt not on t such that
W(t, t + u,) =
t + l1 ,
Jr
e- 2t dt =0.5e- 2t(l- e- 211,) > a I (uc;)
for alIt> O.
Example 5
Consider the one-dimensional dynamical equation
b(t)u
with b(t) defined as in Figure 5-5. The dynamical equation is not instantaneous
ly controllable in the interval (-1, 1). However, it is uniformly controllable
in ( 00, oo). This can be easily verified by choosing U c = 5.
A remark is in order concerning controllability, differential controllability,

and uniform controllability. If a dynamical equatipn is differentially con
trollable, a state can be transferred to any other state in an arbitrarily small
interval of time. However, the magnitude of the input may become very large;
in the extreme case, a delta-function input is required. If a dynamical equation
is merely controllable, the transfer of the states may take a very long interval of
time. However, if it is uniformly controllable, the transfer of the states can be
Theorem 5-7
The n-dimensional linear time-inva

if and only if any of the following ec
1. All rows of e-AtB (and conseqt:

[0, (0) over C, the field of comp
1', All rows of (sI _A)-IB are line.
6
Although all the entries of A and Bare real I

of the field of complex numbers.
AR DYNAMICAL EQUATIONS
CONTROLLABILITY OF LINEAR DYNAMICAL EQUATIONS
183
b(t)
y controllable if and only if there

1 U e such
that
S Ct2(U e)I
~*(t
+Ue, t) ~Ctiuc)I
-3
-2
-1
.tion matrix and W is as defined
is a positive definite matrix (see

I that the transfer of the states
Dcept of uniform controllability
:rol systems. See References 56
controllability both imply con
lIability neither implies nor is
Figure 5-5 The function b(t).
achieved in the time interval U c ; moreover, the magnitude of the control input
will not be arbitrarily large [see (5-14) and note that the input is proportional to
W 1]. In optimal control theory, the condition of uniform controllability is
sometimes required to ensure the stability of an optimal control system.
Time-invariant case. In this subsection, we study the controllability of the
n-dimensional linear time-invariant state equation
FE:
x = Ax +Bu
(5-23)
1equation
lical equation is instantaneously

lical equation is not uniformly
:nds on Ue but not on t such that
tion
al equation is not instantaneous

ver, it is uniformly controllable
Dosing U c 5.
lility, differential controllability,

equation is differentially con
ler state in an arbitrarily small
le input may become very large;
luired. If a dynamical equation
may take a very long interval of
the transfer of the states can be
where x is the n x 1 state vector, u is the p x 1 input vector: A and Bare n x n

and n x p real constant matrices, respectively. For time-invariant dynamical
equations, the time interval of interest is from the present time to infinity;
that is. [0. (0).
The condition for a time-var.ying state equation to be controllable at to is
that there exists a finite t 1 such that all rows of cb(t 0, t )B(t) are linearly independent,
in t, on [to, t l ]. In the time-invariant case, we have (,1)(10' t)B(t) eA(to-t)B.
As discussed in Chapter 4. all elements of e Alto -1)8 are linear combinations of
terms of form Ike;'/; hence they are analytic on [0. 00) (see Appendix B). Con
sequently, if its rows are linearly independent on [0. (0). they are linearly
independent on [1o, tlJ for any to and any 11 > to. In other words, if a linear
time-invariant state equation is controllable, it is controllable at every to? 0
and the transfer of any state to any other state can be achieved in any nonzero
time interval. Hence the reference of 10 and t I is often dropped in the con
trollability study of linear time-invariant state equations.
Theorem 5-7
The n-dimensional linear time-invariant state equation in (5-23) is controllable
if and only if any of the following equivalent conditions is satisfied:
1. All rows of e-AtB (and consequently of eAtB) are linearly independent on
[0, (0) over C, the field of complex numbers. 6
1'. All rows of (sI - A)-l B are linearly independent over C.
6
Although all the entries of A and B are real numbers, we have agreed to consider them as elements
DYNAMICAL EQUATIONS
183
b(t)
v controllable if and onl y if there
(J'c such that
:;; (l2((J' c)1
,*(t
+(J'c' t)~(li(J'c)1
-3
-2
-1
lion matrix and W is as defined
ls a positive definite matrix {see

that the transfer of the states
!lcept of uniform controllability
rol systems. See References 56
:ontrollability both imply con

lability neither implies nor is
Figure 5-5 The function b(t}.
achieved in the time interval (J'c; moreover, the magnitude of the control input
will not be arbitrarily large [see (5-14) and note that the input is proportional to
W- 1 ]. In optimal control theory, the condition of uniform controllability is
sometimes required to ensure the stability of an optimal control system.
Time-invariant case. In this subsection, we study the controllability of the
n-dimensional1inear time-invariant state equation
FE:
x = Ax +Bu
(5-23)
equation
cal equation is instantaneously

ical equation is not uniformly
lds on (J'c bilt not on t such that
lon
I equation is not instantaneous

er, it is uniformly controllable
osing (J'c = 5.
Iity, differential controllability,

~quation is differentially con
~r state in an arbitrarily small
: input may become very large;
lired. If a dynamical equation
(lay take a very long interval of
he transfer of the states can be
where x is the n x 1 state vector, u is the p x 1 input vector; A and Bare n x n

and n x p real constant matrices, respectively. For time-invariant dynamical
equations. the time interval of interest is from the present time to infinity:
that is. [0. (0).
The condition for a time-varying state equation to be controllable at to is
that there exists a finite t 1 such that all rows of (1)(10' t)B(t) are linearly independent,
in t. on [to. t l ]. In the time-invariant case. we have $(10' t)B(t)=eA(to-t)B.
As discussed in Chapter 4. all elements of eA(to-tlB are linear combinations of
terms of form tke;'t; hence they are analytic on [0. XJ) (see Appendix B). Con
sequently. if its rows are linearly independent on [0. cJ). they are linearly
independent on [to. t l ] for any to and any tl > to. In other words. if a linear
time-invariant state equation is controllable. it is controllable at every to ;:::0
and the transfer of any state to any other state can be achieved in any nonzero
time interval. Hence the reference of to and t I is often dropped in the con
trollability study of linear time-invariant state equations.
Theorem 5-7
The n-dimensional linear time-invariant state equation in (5-23) is controllable
if and only if any of the following equivalent conditions is satisfied:
1. All rows of e-AtB (and consequently of eAtB) are linearly independent on

[0, (0) over C, the field of complex numbers. 6
I'. All rows of (sl A)-l B are linearly independent over C.
6
Although all the entries of A and B are real numbers, we have agreed to consider them as elements
184
COl'
2. The controllability grammian
Example 6
W,,~ J~ e"BB'e A ', dT
Consider the inverted pendulL

equation is developed in (3-42)
2g(A1 +m)j(21W +m)/ 5,2/(21
(3-42) becomes
is nonsingular for any t > 0. 7

3. The n x (np) controllability matrix
U~[B:
AB: A2B : ... : A"-IB]
(5-24 )
has rank n.
4. For every eigenvalue Aof A (and consequently for every). in C), the n x (n
complex matrix [)J - A : B] has rank n. 8
*=[1
+ p)
We compute
-1
Proof
The equivalence of statements 1 and 2 follows directly from Theorems 5-1 and
5-4. Since the entries of e- AtB are analytic functions, Theorem 5-3 implies that
the rows of e - AtB are linearly independent on [0, 00) if and only if
p[ e- AtB : - e- AtAB : . . . : ( 1)"- 1 e - AtA IB : ...]

11-
1)" - 1 A" - IB : (
1)nAnB : ...] = n
From the Cayley~-Hamilton theorem, we know that Am with m ~n can be written

as a linear combination of I, A, ... , A"-I; hence the columns of AmB with
m ~n are linearly dependent on the columns of B, AB, ... , Ail-lB. Conse
quently,
p[B: -AB: ... : (_l)"-IAn- t B : .. 0] p[B: -AB: .00: ( l)"-IA"- IB]
Since changing the sign will not change the linear independence, we conclude
that the rows of e- AtB are linearly independent if and only if pCB : AB : 00 0
: An - I B] n. This proves the equivalence of statements 1 and 3. In the
foregoing argument we also proved that the rows of e - AtB are linearly inde
pendent if and only if the rows of eAtB are linearly independent on [0, 00) over
the field of complex numbers. Next we show the equivalence of statements 1
and 1'. Taking the Laplace transform of eAtB, we have
.se[eAtB]
(sI - A)-l B
Since the Laplace transform is a one-to-one linear operator, if the rows of

eAtB are linearly independent on [0, 00) over the field of complex numbers, so
are the rows of (sI - A)-l B, and vice versa.
The proof of statement 4 will be postponed to Section 5-5 (page 206). Q.E.D.
The matrix is in fact positive definite. See Problem E-11.

8
This condition implies that lsI
A) and B are left coprime. See Appendix G.
[B AB
A2B
for any tin [0,00). Let t =0; then the equation reduces to
p[B: - AB: ..
This matrix can be readily shO\

Th us, if x 3 = e is different from,
to bring it back to zero. In fact,
derivatives back to zero. Thh
balancing a broom on our han(
. Example 7
Consider the platform system st

of the platform is zero and the
systems. The spring constants
coefficients are assumed to be 2 a
X2
+x2=u,or
[~J=[
This is the state-variable descrip

Now if the initial displacem(
platform will oscillate, and it ~
platform to come to rest. No~
and x 2 (0) = - 1, is it possible to
seconds? The answer does not
applied to the two spring system
For the state equation in (5-~
p[B: AB
hence the state equation is contr
)YNAMICAL EQUATIONS
185
Example 6
Consider the inverted pendulum system studied in Figure 3-15. Its dynamical
equation is developed in (3-42). For convenience, we assume 2mg/(2M +m) = 1,
2g(M +m)j(2M +m)l 5,2f(2M +m)= 1, and 1/(2M +m)I=2. Then Equation
(3-42) becomes
dt
(5-24)
)r every}, in C), the n x (n
+ p)
=[1
1
0
0
0
0
-1
0
5
llx+
We compute
~ctly
from Theorems 5-1 and

fns, Theorem 5-3 implies th~t
00 ) if and only if
j4.n-
B: ... J=n
~duces
to
,Am with m ;? n can be written

:e the columns of AmB with
B, AB, ... , An-lB. Conse
U = [B
AB
A2B
A 3 BJ
[J}
[j
[1
1
0
-2
0
-10
0 OJx
(5-25)
-Ill
This matrix can be readily shown to have rank 4. Hence (5-25) is controllable.
Thus, if X3 = 8 is different from zero by a small amount, a control u can be found
to bring it back to zero. In fact, a control exists to bring x I = y, X3 = 8, and their
derivatives back to zero. This is certainly consistent with our experience of
balancing a broom on our hand.
, Example 7
-AB:"': (-l)n-1An-1BJ
. independence, we conc1 ude
f and only if p[B : AB : ...
tatements 1 and 3. In the
; of e- AtB are linearly inde
independent on [0, (0) over
equivalence of statements 1
have
!ar operator, if the rows of

ield of complex numbers. so
~tion 5-5
(page 206). Q.E.D.
Consider the platform system shown in Figure 5-2. It is assumed that the mass
of the platform is zero and the force is equally divided among the two spring
systems. The spring constants are assumed to be 1, and the viscous friction
coefficients are assumed to be 2 and 1 as shown. Then we have Xl +2Xl = u and
X2
=U, or
(5-26)
This is the state-variable description of the system.

Now if the initial displacements XI(O) and X2(O) are different from zero, the
platform will oscillate, and it will take, theoretically, an infinite time for the
platform to come to rest. Now we pose the following problem: If Xl (0) = 10
and X2(0) = - 1, is it possible to apply a force to bring the platform to rest in 2
seconds? The answer does not seem to be obvious because the same force is
applied to the two spring systems.
For the state equation in (5-26), we compute
p[D: AD]
.ee Appendix G.
p [~.5
=~25J =2
hence the state equation is controllable. Consequently, the displacements can
186
be brought to zero in 2 seconds by applying a proper input u. Using Equations

(5-13) and (5-14), we have
(J[0.5J [0.5
O Sf
W(O, 2) -_
J.2
[e
,f
J .r _ -
[1.6
(
CON
and
Ul(t)= -[0.5
1]
eo.S
[
O Sf
1] [e
,f
d -
6:B
6.33J
27
for t in [0, 2]. If a force of the

at t = 2. The behavior of Xf, x
lines in Figure 5-6.
u(t)
In" Figure 5-6 we also plot ~

x1(0) = 10 and x 2 (0) = -1 to Zl
in transferring x(O) to zero, the s
of the input. If no restriction
x(O) to zero in an arbitrarily sn
the input may become very lafJ
imposed, we might not be ablf
interval of time. For example
might not be able to transfer x((
40
Example 8
Consider again the platform syst

the viscous friction coefficient all
all equal to 1. Then the state-v
[~J=
Clearly we have
pU=
and the state equation is not con

an input to transfer x(O) to the zc
X2(0), no input can transfer both
X2 (t)
* Controllability
10
Let A and B be n x nand n x

2
oI
4
~
-s
Figure 5-6
indices
Behavior of Xl(t) and X2(t) and the waveform of u.
U k =[B:AB:
It consists of k + 1 block column

The matrix U ~ Un - 1 is the cont
then Un - 1 has a rank of n and con:
Note that there are a total of np co
ways to choose these n linearly ind
ing the most natural and also the m
Let b, ; = 1, 2, ... ,p, be the ith coli
lper input u. Using Equations
6.33J
0e Jdr [1.6
6.33 27
T
187
DYNAMICAL EQUATIONS
and
u,(t)=
-[0.5 It:" ~,J W-'(O, 2) [ ~~J = -44.le5< +20.7e'
for t in [0,2]. If a force of the form Ul is applied, the platform will come to rest
at t = 2. The behavior of Xb X2 and of the input Ul are plotted by using solid
lines in Figure 5-6.

In Figure 5-6 we also plot by using dotted lines the input U2(t) that transfers
x1(0) = 10 and X2(0) = 1 to zero in 4 seconds. We see from Figure 5-6 that,
in transferring x(O) to zero, the smaller the time interval the larger the magnitude
of the input. If no restriction is imposed on the input u, then we can transfer
x(O) to zero in an arbitrarily small interval of time; however, the magnitude of
the input may become very large. If some restriction on the magnitude of u is
imposed, we might not be able to transfer x(O) to zero in an arbitrarily small
interval of time. For example, if we require !u(t)!.::s; 5 in Example 5, then we
might not be able to transfer x(O) to zero in less than 4 seconds.
Example 8
Consider again the platform system shown in Figure 5-2. Now it is assumed that
the viscous friction coefficient and the spring constant of both spring systems are
all equal to 1. Then the state-variable description of the platform system is
Clearly we have
and the state equation is not controllable. If Xl (0) = x 2(0), it is possible to find
an input to transfer x(O) to the zero state in a finite time. However, if Xl(O) i=
X2(0), no input can transfer both Xl(O) and X2(0) to zero in a finite time.
*Controllability indices
Let A and B be n x nand n x p constant matrices. Define
k =0,1,2, ...
~form
of u.
(5-27)
It consists of k + 1 block columns of the form AiB and is of order n x (k + l)p.

The matrix U ~ Un _ 1 is the controllability matrix. If {A, B} is controllable,
then U n - 1 has a rank ofn and consequently has n linearly independent columns.
Note that there are a total ofnp columns in U n - 1 ; hence there are many possible
ways to choose these n linearly independent columns. We discuss in the follow
ing the most natural and also the most important way of choosing these columns.
Let b, i 1,2, ... ,p, be the ith column of B. Then the matrix Uk can be written
188
CONTROLLABILITY AND OBSERV ABILITY OF LINEAR DYNAMICAL EQUATIONS
explicitly as
Uk=[b l
Ab l
bp
b2
A2bl
Ab p
I. .
A 2 bp
ro
rl
: Akb l
Akb p]
(5-28)
"---...'V,.-----'
rk (no. of dependent columns)
Now we search linearly independent columns of Uk in order from left to right;

that is, if a column can be written as a linear combination of its left-hand-side
columns, the column is linearly dependent; otherwise, it is linearly independent.
This process of searching can be carried out by using, for example, the column
searching algorithm discussed in Appendix A. Let rj be the number of linearly
dependent columns in AiB, for i =0,1, ... ,k. IfB has a full column rank, then
ro O. We note that if a column, say Ab 2 , is linearly dependent on its left-hand
side columns, then all Ai b 2 , with j = 2, 3, ... , will be linearly dependent on its
left-hand side columns. Indeed, if
We see that A b 2 is a linear combination of its left-hand-side columns. Pro

ceeding similarly, we can show that Ai b 2 , for j 3,4, ... , are linearly dependent
on its left-hand-side columns. Because of this property, we have
0'::;rO'::;r1'::;r2'::;'"
.::;p
Since there are at most n linearly independent columns in U 00' there exists an
integer J1- such that
0'::;rO'::;r1'::;'"
and
rJl
rJl+l
'::;rJl-I
(5-29a)
(5-29b)
<P
Corollary 5-7
The state equation FE is contn
Un - Ii =
Ab 2 CX l b l +cx 2b 2 + ... +cxpb p +cx p+ IAb 1

A 2b 2 = cx 1 Ab 1 +cx2Ab2 + . " +cx~bp +cx p+ lA 2b 1
then
Hence, all columns of A~ ~

columns. Consequently, we hi
Uk must increase by at least 1
increase. Since the rank of Ua
to add at most n - p number
J1- - 1 .::; n - p or J1-'::; n - p + 1.
yields the upper bound in (5-31
In 'order to have rank n, t1
rows. Hence, we need J1-p;:::n 0
(5-31 ).
The degree ii of the minimal
min (ii, n - p + 1) in (5-31) is (
Theorem 5-7, we have immedia
where p is the rank of B, has

singular.
Example 9
Consider the satellite system stt
equation is developed in (3-51).
the control of
i 1,2, 3, 4, b~
are entirely independent; hence, ,
of (3-51):
1
Equivalently, J1- is the integer such that
pUO<pU l < .. , <pU Jl - 1 =pUJl =pUJl + 1 ='"
(5-30)
In other words, the rank of Uk increases monotonically until k reaches J1- 1;

thereafter, all P columns of AiB will be linearly dependent of their left-hand-side
columns. Hence the controllability of {A, B} can be checked from U Jl - 1 and J1
is called the controllability index. We claim that the controllability index
satisfies the inequalities
~ .::; J1- .::; min(ii, n - p + 1)
(5-31 )
where n is the degree of the minimal polynomial of A and

From the definition of minimal polynomial, we have'
An =cx 1An-1 +cx 2 A n- 2 + ... +cxnI
for some
CXh
x=[l
o
-2
1 0
y [0 0
where we have assumed, for sim

bars. The rank of B is 2; hence,
the matrix
p is the rank of B.
[.8 : AB : A2B]
=[
which implies
AnB =cx 1 An- 1 B +cx2Aii-2B + ... + cxnB
has rank 4. This is the case; hen
LINEAR INDEPENDENCE OF TIME FUNCTIONS
l DYNAMICAL EQUATIONS
,]
(5-28)
mns)
Uk in order from left to right;

nbination of its left-hand-side
wise, it is linearly independent.
sing, for example, the column
,et ri be the number of linearly
B has a full column rank, then
rly dependent on its left-hand
II be linearly dependent on its
Hence, all columns of AiiB are linearly dependent on their left-hand-side

columns. Consequently, we have J.l =:; ii. From (5-30), we see that the rank of
Uk must increase by at least 1 as k increased by 1; otherwise, it will cease to
increase. Since the rank of U oo is at most n, if the rank of B is p, it is sufficient
t
.
to add at most n p number of AlB. Hence, we conclude from (5-30) that
J.l - 1 =:; n - p or J.l =:; n p + 1. The combihation of J.l =:; ii and J.l =:; n - p + 1
yields the upper bound in (5-31).
In order to have rank n, the matrix U/l.- I must have more columns than
rows. Hence, we need J.lp 2n or J.l ?:.n/p. This establishes the lower bound in
(5-31).
The degree i1 of the minimal polynomial of A is not easy to compute; hence
min (ii, n - p + 1) in (5-31) is often replaced by n - p + 1. From (5-30) and
Theorem 5-7, we have immediately the following corollary.
Corollary 5-7
The state equation FE is controllable if and only if the n x (n
[B
Un-Ii
+a p + 1 Ab l
.Abp +a p + lA 2 b 1
left-hand-side columns. Pro
,4, ... ,are linearly dependent
operty, we have
-s;p
olumns in U em there exists an
(5-29 a )
(5-29b)
<p
1 of A and
,ave
(5-31 )
p is
...
An-PB]
Example 9
Consider the satellite system studied in Figure 3-17. Its linearized dynamical
equation is developed in (3-51). As can be seen from the dotted lines in (3-51),
the control of Xi' i 1, 2, 3, 4, by UI and li2 and the control of Xs and X6 by
are entirely independent; hence, we may consider only the following subequation
of (3-51):
[l
(5-30)
t 1)
AB
p + l)p matrix
where p is the rank of B, has rank n, or the n x n matrix Un _pU:_ p is nonsingular.
o
o
onically until k reaches J.l 1;

pendent of their left-hand-side
1 be checked from U /l. - 1 and J.l
:hat the controllability index
189
-2
(5-32a)
=[~ ~ ~ ~J x
(5-32b)
where we have assumed, for SimpliCity, W = m = ro 1 and dropped the over

bars. The rank of B is 2; hence, the state equation is controllable if and only if
the matrix
the rank of B.
[B : AB : A2B]
=[l
o
O'
O'
0
O.. 0
1 : -2
2
1 : -2
O 0
J]
has rank 4. This is the case; hence, the state equation in (5-32) is controllable.
190
the controllability indices of ~

1, .. ,1l}.
For this problem, we can readily show
pU O =2<pU l =4
pU 2 =pU 3 = ...
hence, the controllability index of (5-32a) is 2.
We study more the controllability matrix U. We assume that the linearly

independent columns of U in order from left to right have been found. We
now rearrange these independent columns as
bi> Ab l , ... , AJl1 lb l , b 2 , Ab 2 ,
... ,
AJl2- l bz, ... , b p, Ab p,"" AJlp-lb p
Theorem 5-8
The set of the controllability in
transformation and any orderi
Proof
Define
The integer Ili is the number of linearly independent columns associated with b i
in the set, or is the length of the chain associated with bi' Clearly we have
where A=PAP-l, B PB an
easily verified that
Il =max{llb 1l2, ... ,Il p }
and III + 112 + ... + II p S n. The equality holds if {A, B} is controllable. The
set {Ill' 1l2, .. ,Il p } will be called the controllability indices of {A, B}.
Now we shan establish the relationship between the controllability indices
and the r/s defined in (5-28). In order to visualize the relationship, we use an
example. We assume p = 4, III = 3, 112 1, 113 5, and 114 3. These indepen
dent columns are arranged in a crate diagram as shown in Figure 5-7. The
(i,j)th cell represents the column Ai-lb j . A column which is linearly indepen
dent of its left-hand-side columns in (5-28) is denoted by "x"; otherwise denoted
by "0". The search of linearly independent columns in (5-28) from left to right
is equivalent to the search from left to right in each row and then to the next
row in Figure 5-7. Hence, the number of zeros in the ith row of Figure 5-7 is
equal to r i - l as shown. From the crate diagram, we can deduce that ri is equal
to the number of {b k , k 1, 2, ... , p} with controllability indices equal to or
smaller than i. Hence, we conclude that
no. of {b k , k = 1, 2, ... ,p} with controllability index i
rj - ri-l
Uk =
(5-33)
with r-l~O. For example, rl-rO 1, and b 2 has controllability index 1;

1'1
0 and no b i has controllability index 2; 1'3 - r2 2, and b l and b 4 have
controllability index 3; 1'5 I'4 = 1, and b 3 has controllability index 5. Hence,
Uk=F
which implies
pUk=P
Hence, the rj defined in (5-2~
indices are invariant under an)
The rearrangement of the c
where M is a p x p elementary
forward to verify that
irk~[D
where diag {M, M, ... , M} cor

singular. Hence, we have
pU k
r2 -
Ili
bi
b2 b 3
b4
0
0
0
0
0
x
x
x
x
0
5
x
x
x
0
0
0
3
ro
rl
1'2
=1
r3 =3
r4 3
r5 4
Figure 57 Crate diagram of Ai-1b j .
: AD : ...
pl
Consequently, we conclude tha

the ordering of the columns of:
Now we discuss a different

umns in U =[B AB '" All
b l , Ab 1 , A 2 b b
... ,
An-lb l ; b;
and then search its linearly in<

In terms of the crate diagram i
are searched in order from top t
column and so forth. Let
b l , Ab l , ... , Aill- 1 b 1 ; b 2 , AI
LINEAR INDEPENDENCE OF TIME FUNCTIONS
DYNAMICAL EQUATIONS
191
the controllability indices of {A, B} are uniquely determinable from {rj' i =0,
1, ... , 11}.
=pU 3 =" .
We assume that the linearly

right have been found. We
Theorem 5-8
I
The set of the controllability indices of {A, B} is invariant under any equivalence
transformation and any ordering of the columns of B.
Proof
, .. , bp , Abp ,
. . , AlJ.p-l
bp
nt columns associated with b i

lth bi' Clearly we have
!p}
f {A, B} is controllable. The

ity indices of {A, B}.
~en the controllability indices
ze the relationship, we use an
, and 114 3. These indepen
LS shown in Figure 5-7. The
mn which is linearly indepen
ted by "x" ; otherwise denoted
lUS in (5-28) from left to right
ach row and then to the next
in the ith row of Figure 5-7 is
we can deduce that ri is equal
~ol1abi1ity indices equal to or
controllability index i
(5-33)
: has controllability index 1;

3 - r2
2, and b I and b 4 have
mtrollability index 5. Hence,
Define
Uk [D: AD : ... : AkDJ

where A PAP - 1, D = PB and P is any nonsingular matrix. Then it can be
easily verified that
for k
0, 1, 2, ...
which implies
for k = 0, 1, 2, ...
Hence, the rj defined in (5-28) and, consequently, the set of controllability
indices are invariant under any equivalence transformation.
The rearrangement of the columns of B can be represented by
D=BM
where M is a p x p elementary matrix and is nonsingular. Again it is straight

Jorward to verify that
ih~ [D
: AD : ... : AkBJ = Uk diag{M, M, ... , M}

where diag {M, M, ... ,M} consists of k + 1 number of M, and is clearly non
singular. Hence, we have
for k = 0, 1, 2, ...
Consequently, we conclude that the controllability indices are independent of
the ordering of the columns of B.
Q.E.D.
Now we discuss a different method of searching linearly independent col
umns in U [B AB ... An-l BJ. We first rearrange the columns of U as
b I , Ab b A 2 b 1 ,
... ,
An- 1 b 1 ; b 2 , Ab 2 ,
... ,
A n- 1 b 2 ;
. ;
b p , Ab p , " " An- 1 b p

(5-34)
and then search its linearly independent columns in order from left to right.
In terms of the crate diagram in Figure 5-7, the linearly independent columns
are searched in order from top to bottom in the first column, then in the second
column and so forth. Let
192
be the resulting linearly independent columns. If {A, B} is controllable, we have
Consider the n-dimensiona
iiI +Jiz + ... +Jip=n. The lengths of these chains are {Jib Jiz,, Ji p}.
Unlike the controllability indices, these lengths of chains depend highly on the
ordering of {bj, i = 1, 2, ... ,p}.
E:
where A, B, C, and E are n x n,

continuous functions of t defin
Example 10
Consider the state equation in (5-32). lfwe search linearly independent columns
from left to right in
b l , Ab l , A 2 b!> A 3 b 1 ; b 2 , Ab 2 , AZb z, A 3 b 2
the resulting linearly independent columns are

b l , Ab b A 2 b l , b 2
Its lengths are {3, I}. If we search from left to right in

b 2 , Ab 2 , A 2 b z, A 3 b z ; b l , Ab b A 2 b!> A 3 b I
the resulting linearly independent columns are

bz, Ab 2 , A 2 b 2 , A 3 b z
Its lengths are {4, O}. The lengths are indeed different for different ordering of
b l and b 2 The controllability indices of (5-32) can be computed as {2, 2} and
are independent of the ordering of b i and b 2
Definition 5-5
The dynamical equation E is

if there exists a finite t I > to suc
of the input u[to. td and the out}
to determine the state xo. Ott
unobservable at to.
Example 1
Consider the network shown in

the initial voltage across the capi
the output is identically zero. ,
cally zero), but we are not able to
hence the system, or more preci:
system, is not observable at any
Example 2
Let ii =max{Jib Ji2' ... , Ji p}. It is clear that Ji can never be smaller than {L,
the controllability index. Since {L =max{{Lt. {Lz, ... , {Lp}, we may conclude
that {L is the smallest possible maximum length of chains obtainable in any
search of linearly independent columns of U. This controllability index will
play an important role in the design of state feedback in Chapter 7 and the design
of compensators in Chapter 9.
Consider the network shown in F

reduces to the one shown in Fig
+A
u(rv)
5-4
Observability of Linear Dynamical Equations
Time-varying case. The concept of observability is dual to that of control

lability. Roughly speaking, controllability studies the possibility of steering
the state from the input; observability studies the possibility of estimating the
state from the output. If a dynamical equation is controllable, all the modes of
the equation can be excited from the input; if a dynamical equation is observ
able, all the modes of the equation can be observed at the output. These two
concepts are defined under the assumption that we hC}ve the complete knowledge
of a dynamical equation; that is, the matrices A, B, C, and E are known before
hand. Hence, the problem of observability is different from the problem of
realization or identification. The problem of identification is a problem of
estimating the matrices A, B, C, and E from the information collected at the
input and output terminals.
In~WH
IF
.L
Inc>
~-.....
105-j
----.J-
Figure 5-8 An unobservable netwo

IH
In? I~Jl
-L..
..
(a)
Figure 5-9 An unobservable networ
l DYNAMICAL EQUATIONS
193
OBSERVABlLITY OF LINEAR DYNAMICAL EQUATIONS
{A, B} is controllable, we have

chains are {.u 1, .u 2, . . ., .up}.
)f chains depend highly on the
Consider the n-dimensional linear dynamical equation

E:
x
Y
A(t)x(t) + B(t)u(t)
C(t)x(t) + E(t)u(t)
(5-35a)
(5-35b)
where A, B, C, and E are n x n, n x p, q x n, and q x p matrices whose entries are

continuous functions of t defined over (- 00, 00).
llinearly independent columns
The dynamical equation E is said to be (completely state) observable at to

if there exists a finite t 1 > to such that for any state Xo at time to, the knowledge
of the input u[to,td and the output Y[to,td over the time interval [to, t 1] suffices
to determine the state Xo. Otherwise, the dynamical equation E is said to be
unobservable at to
ight in
Example 1
A2bb A3 b 1
fferent for different ordering of

can be computed as {2, 2} and
.u can never be smaller than 11,

~2'
... ,
Definition 5-5
II p }, we may conclude
h of chains obtainable in any

This controllability index will
ack in Chapter 7 and the design
Consider the network shown in Figure 5-8. If the input is zero, no matter what
the initial voltage across the capacitor is, in view of the symmetry of the network,
the output is identically zero. We know the input and output (both are identi
cally zero), but we are not able to determine the initial condition of the capaci~or;
hence the system, or more precisely, the dynamical equation that describes the
system, is not observable at any to.
Example 2
Consider the network shown in Figure 5-9(a). Ifno input is applied, the network
reduces to the one shown in Figure 5-9(b). Clearly the response to the initial
1
1
In
u'"'-'
cal Equations
bility is dual to that of control
lies the possibility of steering
le possibility of estimating the
s controllable, all the modes of
dynamical equation is Qbserv
'Ved at the output. These two
e have the complete knowledge
B, C, and E are known before
different from the problem of
identification is a problem of
e information collected at the
In
Figure 5-8 An unobservable network.

IH
I H
In
1
' - - -_ _----4-----J
F
(a)
Figure 5-9 An unobservable network.
(b)
194
CONTROLLABILITY AND OBSERV ABILITY OF LINEAR DYNAMICAL EQUA nONS
current in the inductor can never appear at the output terminal. Therefore,
there is no way of determining the initial current in the inductor from the input
and the output terminals. Hence the system or its dynamical equation is not
observable at any to
Ifto <1>(t, r)B(r)u(r)dr +E(t)u(t)
C( t)<l>(t. t o)x(t 0)
y(t)~y(t) -
where
CU)
(5-37)
Itto <1>(t. r)8(t)u(r) lit - E(t)uU)
(5-38)
is a known function. Consequently, the observability problem is a problem of

determining x(t o) in (5-37) with the knowledge of y. C, and <1>(t. to). Note that
the estimated state x(to) is the state not at time t, but at time to. However, if
x(t o) is known, the state after to can be computed from
x(t) = <1>(t, to)x(to) +
Itto <1>(t, t )B( t)u( t) dr
(5-39)
The dynamical equation E is observable at to if and only if there exists a finite

t 1> to such that the n columns of the q x n matrix function C(')<l>(', to) are linearly
independent on [to, t 1].
Proof
Sufficiency: Multiplying <1>*(t, to)C*(t) on both sides of (5-37) and integrating

from to to t 1, we obtain
to
<1>*(t, to)C*(t)y(t) dt
[L
<1>*(1, 10lC*( t )C(t )<1>(1, (0)
dlJ Xo ~ V(l o, ( 1)xo

(5-40)
tl
where
V(to, t d~
<1>*(t, to)C*(t)C(t)<l>(t, to) dt
then
Hence the initial statex(t o) IX C

tion that E is observable. Th
t 1> to such that the columns 0
We see from this theorem tt

tion depends only on C(t) and j
can also be ded uced from Defi]
servability study, it is sometim
x = A(t)x. y = C(t)x.
The controllability of a dy
independence of the rows of <1>(t
by the linear independence of
between these two concepts is e
Theorem 5-10 (Theorem of dl
Theorem 5-9
tl
IX;
y(t)=C(t)fj
(5-36)
where <1>(t. r) is the state transition matrix of x = A(t)x. In the study of observ
ability. the output y and the input u are assumed to be known. the initial state
x(t o ) is the only unknown; hence (5-36) can be written as
y(t)
C(t}fj
Let us choose x(t o)
The response of the dynamical equation (5-35) is given by

y(t)=C(t)l>(t. to)xU o) +C(t)
Prove by contradiction. Sup

t 1> to such that the columns (
Then there exists an n x 1 non:
(5-41 )
to
From Theorem 5-1 and the assumption that all the columns of C(')<l>(', to) are
linearly independent on [to, t 1 ], we conclude that N(to, td is nonsingular.
Hence, from (5-40) we have
Consider the dynamical equati4

defined by
E*:
where A*, B*, C*, and E* are thl

E in E. The equation E is COl
equation E* is observable (contI
Proof
From Theorem 5-4, the dynamic

rows of <1>(t o, t)B(t) are linearly i
5-9, the dynamical equation E'
B*(t)<1>a(t, to) are linearly indepeOl
[B*(t)<1>a(t, t o)]* <1>:(t, t o}8(t} ar
I>a is the state transition matrix of
<1>(to, t) (see Problem 4-8); hence J
tl
Xo = V- 1 (to, td
<1>*(t, to)C*(t)y(t) dt
(5-42)
to
Thus, if the function y[to, ttl is known, Xo can be computed from (5-42). Necessity:
We list in the follOWing, for 01

tions 5-6 to 5-8, which are dual'
5-4 for controllability. Theorem
OBSERVABILITY OF LINEAR DYNAMICAL EQUATIONS
IYNAMICAL EQUATIONS
mtput terminal. Therefore,

the inductor from the input
s dynamical equation is not
Prove by contradiction. Suppose E is observable at to, but there exists no

tl > to such that the columns of C(')(f)(', to) are linearly independent on [to, tIl
Then there exists an n x 1 nonzero constant vector (I such that
r)u(r) elr
+ E(t)u(t)
(5-37)
(5-38)
lity problem is a problem of

C and (f)(t, to). Note that
but at time to. However, if
i,
from
r)u(r) elr
(I; then
y(t) = C(t}(f)(t, to)(I
In the study of observ

o be known, the initial state
.ten as
E(t)u (t )
for all t> to
for all t> to
(5-36)
[t)x.
:lr
Qt)(f)(t, to)(I =0
Let us choose x(to)
is given by
195
(5-39)
Hence the initial state x(t o) (I cannot be detected. This contradicts the assump
tion that E is observable. Therefore, if E is observable, there exists a finite
tl > to such that the columns ofC('}(f)(" to) are linearly independent on [to, tIl
Q.E.D.
We see from this theorem that the observability of a linear dynamical equa
tion depends only on C(t) and (f)(t, to) or, equivalently, only on C and A. This
can also be deduced from Definition 5-5 by choosing u O. Hence in the ob
servability study, it is sometimes convenient to assume u 0 and study only
x = A(t)x, y C(t)x.
The controllability of a dynamical equation is determined by the linear
independence of the rows of (f)(to, ')B(,), whereas the observability is determined
by the linear independence of the columns of C(')(f)(', to). The relationship
between these two concepts is established in the following theorem.
Theorem 5-10 (Theorem of duality)
d only if there exists a finite

lction C(')(f)(', to) are linearly
es of (5-37) and integrating
~t, to) dtJxo~ V(to, ttl"o

(5-40)
(5-41 )
Consider the dynamical equation E in (5-35) and the dynamical equation E*

defined by
E*:
i
"(
A*(t)z +C*(t)v
B*(t)z + E*(t)v
(5-43a)
(5-43b)
where A*, B*, C*, and E* are the complex conjugate transposes of A, B, C, and
E in E. The equation E is controllable: (observable) at to if and only if the
equation E* is observable (controllable) at to.
Proof
From Theorem 5-4, the dynamical equation E is controllable if and only if the
rows of (f)(t o, t)B(t) are linearly independent, in t, on [to, tIl From Theorem
5-9, the dynamical equation E* is observable if and only if the columns of
B*(t)4it, to) are linearly independent, in t, on [to, tlJ, or equivalently, the rows of
[B*(t)4it, to)J* = 4:(t, to)B(t) are linearly independent, in t, on [to, t IJ, where
e columns of C(')(f)(', to) are

at V(t o, t 1) is nonsingular.
4a is the state transition matrix ofi - A *(t)z. It is easy to show that 4:(t, to)
(f)(to, t) (see Problem 4-8); hence E is controllable if and only if E* is observable.
Q.E.D.
t)y(t) elt
We list in the following, for observability, Theorems 5-11 to 5-14 and Defini
tions 5-6 to 5-8, which are dual to Theorems 5-5 to 5-8 and Definitions 5-2 to
5-4 for controllability. Theorems 5-11 to 5-14 can be proved either directly or
(5-42)
Ited from (5-42). Necessity:
196
by applying Theorem 5-10 to Theorems 5-5 to 5-8. The interpretations in the

controllability part also apply to the observability part.
Definition 5- 7
Theorem 5-11
( - 00, 00) if and only if
The linear dynamical equati4

Assume that the matrices A(') and C() in the n-dimensional dynamical equation
E are n - 1 times continuously differentiable. Then the dynamical equation E
is observable at to if there exists a finite t 1 > to such that
1
NO(t )
N1(tl)
P
.
=n
[
Nn-1(t 1)
where
N k+ 1 (t) = Nk(t)A(t)
WIth
+ dt
k=0,1,2, ... ,n-1
Nk(t)
No(t) = C(t)
(S-44 )
No(t)
N1(t)
[
p N"-l(t)
where the N/s are as defined i

Definition 5-8
(S-4Sa)
(S-4Sb)
* Differential
The linear dynamical equatior

if and only if there exist a posit
0< Pl(aJI ~V(t, t
0< P3((Jo)I ~4l>*(t,
observabilitv, instantaneous observability,

and uniform observability
for all t, where <I> is the state tl
Differential, instantaneous, and uniform observabilities can be defined by using

the theorem of duality; for example, we may define {A, C} to be differentially
observable if and only if { - A*, C*} is differentially controllable. However,
for ease of reference we shall define them explicitly in the following.
Time-invariance case.
equation
FJ
Definition 5-6
The dynamical equation E is said to be differentially observable at time to if,

for any statex(to) in the state space L, the knowledge of the input and the output
over an arbitrarily small interval of time suffices to determine x(t o).
Theorem 5-12
If the matrices A and C are analytic on ( - 00, 00), then the n-dimensional dy
namical equation E is differentially observable at every t in ( - 00,00) if and only
if, for any fixed to in ( - 00, 00),
No(t o)
N 1 (to)
PI
where A, B, C, and E are n x n

time interval of interest is [0, C1.
time-invariant dynamical equ
to ;;::: 0, and the determination 0
time interval. Hence, the refen
bility study of linear time-inval
Theorem 5-13
The n-dimensional linear time-i

able if and only if any of the fo
I
=n
N n - 1 (to)
*Sections noted with an asterisk may be skipped without loss of continuity.
1. All columns of CeA! are lim

complex numbers.
I'. All columns ofC(sI-Atl
2. The observability grammian
v
is nonsingular for any t > O.
'NAMICAL EQUATIONS
OBSERVABILITY OF LINEAR DYNAMICAL EQUATIONS
The interpretations in the

)art.
197
Definition 5-7
The linear dynamical equation E is said to be instantaneously observable in

( - 00, (0) if and only if
I,
lsional dynamical equation

l the dynamical equation E
that
(5-44 )
p[
~:~:~ ]=n
for all t in (
00,
(0)
Nn~l(t)
where the N/s are as defined in (5-45).

Definition 5-8
1,2, ... ,n
(5-45a)
(5-45b)
Jbility,
The linear dynamical equation E is said to be uniformly observable in ( 00, (0)

if and only if there exist a positive (10 and positive {3i that depends on (10 such that
0< {31((10)1 sV(t, t +(10)s{32((10)1

0< {33((10)1 sf1l*(t, t +(1o)V(t, t +(1o)cl>(t, t +(10) s{34((10)1
for all t, where
lties can be defined by using

;: {A, C} to be differentially
Ily controllable. However,
in the following.
lily observable at time to if,

~ ofthe
input and the output

determine x(to).
then the n-dimensional dy

ery t in 00,(0) if and only
(J)
is the state transition matrix and V is as defined in (5-41).
Time-invariance case.
equation
Consider the linear time-invariant dynamical
FE:
X Ax +Bu
y=Cx +Eu
(5-46a)
(5-46b)
where A, B, C, and E are n x n, n x p, q x n, and q x p constant matrices. The

time interval of interest is [0, (0). Similar to the controllability part, if a linear
time-invariant dynamical equation is observable, it is observable at every
to 2:::0, and the determination of the initial state can be achieved in any nonzero
time intervaL Hence, the reference of to and t 1 is often dropped in the observa
bility study of linear time-invariant dynamical equations.
Theorem 5 -1 3
The n-dimensional linear time-invariant dynamical equation in (5-46) is observ

able if and only if any of the following equivalent conditions is satisfied:
of continuity.
1. All columns of Ce A ! are linearly independent on [0, (0) over C, the field of
complex numbers.
I'. All columns of C(sl - A)-l are linearly independent over C.
2. The observability grammian
w"'~
is nonsingular for any t > 0.
/'C'Ce , dr
198
CANONICAL DECOMPOSITION OF
3. The nq x n observability matrix
and
C
Equivalently, v is the integer s
CA
CA 2
The integer v is called the obs

have
CAn-l
has rank n.
4. For every eigenvalue A of A (and consequently for every Ain C), the (n +q) x n
. complex matrix
[AICA]
has rank n, or equivalently, (sI
Let A and C be n x nand q x n constant matrices. Define
,-
CJ
CA
"0 (no. of dependent rows)
v n
and
CIA}
czA
,.,
+ V2 + . "
+\'q :::;11.
Th
I'k
It consists of k + 1 block rows of the form CA and is of order (k + 1)q x n. The

matrix V = Vn _ 1 is the observability matrix. The q rows of C are denoted by
Cj, i = 1, 2, ... ,q. Let us search linearly independent rows of Vk in order from
top to bottom. Let rj be the number of linearly dependent rows in CA i ,
i =0, 1, ... ,k. Similar to the controllability part, we have
i
'I
1'0 '::;;;1'1 :::;1'2:::;"
:::;q
Since there are at most n linearly independent rows in V

v such that
0'::;;;1'0:::;1'1:::;'"
The set of observability indice

transformation and any orderil
5-5 Canonical Decom~

Dynamical Equation
VI
Theorem 5-14
(5-48)
~;Akl
}
c A
clJA
where ii is the degree of the m;
Consider the matrix Vn - l'

in order from top to bottom h~
independent rows associated ,
the observability indices of {A,
[C~. = _C_q~J
:::;
The dynamical equation FE i

Vn--q, where ij is the rank of C
V:--qVn_-q is nonsingular.
*Observability Indices
V. _
n
q
Corollary 5-13
A) and C are right coprime.
~: 11
pVO<pV 1 < .
(5-47)
Cq
:::;r v - l <q
00'
there exists an integer

(5-49a)
In the remainder of this chap

dynamical equations. Conside
Fj
where A, B, C, and E are n x n

We introduced in the previou
observability. The conditions
able are also derived. A questior
said if the equation is upcontro
shall study this problem. Befol
transformation. Let x Px, w
AMICAL EQUATIONS
CANONICAL DECOMPOSITION OF A LINEAR TIME-INVARIANT DYNAMICAL EQUATION
and
199
(5-49b)
Equivalently, v is the integer such that
pV 0 < PV 1 < .. i < PV v-I
(5-47)
PV v
PV v + 1 =' . .
(5-50)
The integer v is called the observability index of {A, C}. Similar to (5-31), we
have
very A. in C), the (n
n
. (_
-svsmm n,n
+ q) x n
q+ 1)
(5-51 )
where ii is the degree of the minimal polynomial of A and ij is the rank of C.

Corollary 5-13
1t coprime.
The dynamical equation FEin (5-46) is observable if and only if the matrix
V n - q, where ij is the rank of C, is of rank n, or equivalently, the n.x n matrix
V:-qV n _q is nonsingular.
efine
Consider the matrix Vn _ l ' It is assumed that its linearly independent rows
in order from top to bottom have been found. Let Vi be the number of linearly
independent rows associated with C i . The set of {Vi, i = 1,2, ... ,q) is called
the observability indices of {A, C}. Clearly we have
:nt rows)
v = max [ Vi' i
and
VI +V2 + . . . +Vq Sll.
1, 2, ... , q 1
The equality holds if [A,
C: is observable.
Theorem 5-14
(5-48)
The set of observability indices of tA, Cj is invariant under any equivalence

transformation and any ordering of the rows of C.
5-5 Canonical Decomposition of a Linear Time-Invariant

Dynamical Equation
f order (k + 1)q x n. The

ows of C are denoted by
rows of Vk in order from
:lependent rows in CA i ,
have
, 00'
there exists an integer

(5-49a)
In the remainder of this chapter, we study exclusively linear time-invariant

dynamical equations. Consider the dynamical equation
FE:
x=Ax +Bu
y=Cx +Eu
(5-52a)
(5-52b)
where A, B, C, and E are n x n, n x p, q x n, and q x p real constant matrices.

We introduced in the previous sections the concepts of controllability and
observability. The conditions for the equation to be controllable and observ
able are also derived. A question that may be raised at this point is: What can be
said if the equation is uncontrollable and/or unobservable? In this section we
shall study this problem. Before proceeding, we review briefly the equivalence
transformation. Let x = Px, where P is a constant nonsingular matrix. The
200
CONTROLLABILITY AND OBSERVABlLITY OF LINEAR DYNAMICAL EQUATIONS
substitution of x
Px into (5-52) yields
FE:
Proof
Ax +Du
(5-53a)
(5-53b)
If the dynamical equation F1

have
where A = PAp-I, D = PB, C = Cp-I, and E E. The dynamical equations

FE and FE are said to be equivalent, and the matrix P is called an equivalence
transformation. Clearly we have
pU~p[B
y=Cx +Eu
iT~ [D
Let q1, q2, ... , qn 1 be any nl Ii

each i= 1, 2, ... , nb Aqj c.
{qb q2,'" qnJ. (Why?) De
: AD : ... : An-ID] = [PB : PAB :' ... : PAn-IB]

=P[B: AB: ... : An-IB]~PU
P-l~Q~
Since the rank of a matrix does not change after multiplication of a nonsingular
matrix (Theorem 2-7), we have rank U rank U. Consequently FE is con
trollable if and only if FE is controllable. A similar statement holds for the
observability part.
where the last n n i columns

Q is nonsingular. We claim tl
into the form of (5-54). Rec:
x Px we are actually using th
state space. The ith column of
Aqi with respect to {qb qz, ... ,
linearly dependent on the set {~
given in (5-54a). The columns
B with respect to {qb q2, ... ,
{qj, q2" . " q!!.1}; hence B is of
Let U and U be the controlla
we have pU = piT =nl (see The
Theorem 5-15
The controllability and observability of a linear time-invariant dynamical

equation are invariant under any equivalence transformation.
This theorem is in fact a special case of Theorems 5-8 and 5-14. For easy
reference, we have restated it as a theorem.
In the following, c will be used to stand for controllable, cfor uncontrollable,
o for observable, and 0 for unobservable.
V = [Be A)
Theorem 5-16
Consider the n-dimensional linear time-invariant dynamical equation FE.

If the controllability matrix of FE has rank n 1 (where nl < n), then there exists
an equivalence transformation :X = Px, where P is a constant nonsingular
matrix, which transforms FE into
FE:
[~~J
Xc
[Ae0 At
~~. [~:J
+
xe
2J
Y= [Cc C,]
[:;J
[DcJ
0
+Eu
(5-54a)
(5-54b)
Xe Acxe +Dcu
y =Cc:Xc +Eu
Vc
(5-55a)
(5-55b)
is controllable 9 and has the same transfer function matrix as FE.
=[~c
X:
represents the contre
A!B with k '2:nl are linearly de
pV =
n 1 implies pVc n l . Heno

We show now that the dynarr.
FE and FEe' have the same traI
A~..z ]
SI - Ac
sl
[(Sl
Ac
- [SI-X,
0
[Cc Cia
-.Aet
0
-X12
- [(SI
sl-Ai.'
=[Cc C c]
=Ce(sl
is easy to show that if the equation Fit is observable. then its sUbequation Fit, is also observable.
(Try)
9 It
Thus the transfer-function matri
and the ni-dimensional subequation of FE
FEe:
where
Ae)-IBe -t
which is the transfer-function rna
201
Proof
(5-53a)
(5-53b)
The dynamical equations

x P is called an equivalence
:' ... : PAn-1B]

,.: An-lB]~PU
If the dynamical equation FE is not controllable, then from Theorem 5-7 we

have
pU~p[B : AB : ... : An-lB]
Let qb q2' ... , qnl be any nl linearly independent columns of U. Note that for
each i = 1, 2, ... , n l , Aqj can be written as a linear combination of
{ql' q2' .. ,qnJ. (Why?) Define a nonsingular matrix
(5-56)
ltiplication of a nonsingular
Consequently FE is con
lar statement holds for the
time-invariant dynamical
formation.
ms 5-8 and 5-14. For easy
)llable, cfor uncontrollable,
where the last n - nl columns of Q are entirely arbitrary so long as the matrix
Q is nonsingular. We claim that the transformation x Px will transform FE
into the form of (5-54). Recall from Figure 2-5 that in the transformation
x Px we are actually using the columns ofQ~ p-l as new basis vectors of the
state space. The ith column of the new representation A is the representation of
Aqj with respect to {ql' q2,"" qn}. Now the vectors Aqh for i 1,2, ... , nl, are
linearly dependent on the set {q t, q2, ... , qnl} ; hence the matrix A has the form
given in (5-54a). The columns of B are the representations of the columns of
B with respect to {ql' q21 ... ,qn}' Now the columns of B depend only on
{qt, q2,"" q!!J}; hence B is of the form shown in (5-54a)._
Let U and U be the controllability rna trices of FE and FE, respectively. Then
we have pU pU = nl (see Theorem 5-15). It is easy to verify that
- [Bc:: AJic!I " ' ,! A~-l Bc]
U=
0: 0 :I
l dynamical equation FE.

:re n 1 < n), then there exists
is a constant nonsingular
Bc
= [ Uc:! A~lBc!I ' "1L~n-l
c
0:
where
]}n
rows
}(n - n d rows
Uc represents the controllability matrices of FEe- Since columns of
A~B with k :?:nl are linearly dependent on the columns of Uc, the condition
pU
(5-54a)
(5-54b)
nl < n
n1 implies pUc = n 1. Hence the dynamical equation FEc is controllable.

We show now that the dynamical equations FE and FEe' or correspondingly,
and
have the same transfer-function matrix. It is easy to verify that
Sl OAc
[
-A~] [(SI -.Ac)-l

sl - Ac
(sl -Ac)-1A~?(SI1-Ac)-1]
(sl - Ad-
(5-57)
Thus the transfer-function matrix of FE is

-A12]-l[Bc] +E
sl -Ac
0
(5-55a)
(5-55b)
matrix as FE.
lubequation FEe is also observable.
[C, C,] [<Sl
;,)-
(.'II -
Act 1 A 12 (sl
{sl
Ac)-l
= Ce(sl - Ac)-1Bc +E
which is the transfer-function matrix of FEe
Q.E.D.
202
CONTROLLABILn', AND r,BSERVABILITY OF LINEAR DYNAMICAL EQUATIONS
In the equivalence transformation x = Px, the state space L of FE is divided

into two subspaces. One is the n1-dimensional subspace of L, denoted by L 1,
which consists of all the vectors
[~'
and
C Cp-l [
the other is the (n - nIl-dimensional
subspace, which consists of all the vectors
[:J
Hence, the reduced controllab
Since FE, is controllable, all
the vectors Xc in L1 are controllable. Equation (5-54a) shows that the state
variables in Xc are not affected directly by the input u or indirectly through the
state vector Xc; therefore, the state vector is not controllable and is dropped in
the reduced equation (5-55). Thus, if a linear time-invariant dynamical equation
is not controllable, by a proper choice of a basis, the state vector can be decom
posed into two groups: one controllable, the other uncontrollable. By
dropping the uncontrollable state vectors, we may obtain a controllable dy
namical equation of lesser dimension that is zero-state equivalent to the
original equation. See Problems 5-22 and 5-23.
x,
Dual to Theorem
dynamical equations.
G~J
5-16, w
Theorem 5-17
Consider the n-dimensional
If the 0 bservability matrix of 1
valence transformation X
Example 1
Px
[~~
Consider the three-dimensional dynamical equation
Xii
0] [0 1J
x= [01 1
1 0 X+ IOu
o 11
01
y=[1
l]x
(5-58)
The rank of B is 2; therefore, we need to check U 1 = [B : AB] in determining

the controllability of the equation. Since
p[B : AB]
0 1 1 01] = 2 < 3
[o 1 1
p 1 0
01 0:1: 01]
[
o 110
PB~[~
0
0
-1
1 1 0
-:lG ~] [~--~]
This theorem can be readil~

The first n2 rows of P in Theon
the observability matrix of {A,
arbitrary so long as P is nonsi
xii does not appear directly in t
vector xij is not observable and
I
Combining Theorems
theorem.
5-16
The first two columns of Q are the first two linearly independent columns of
U1; the last column of Q is chosen arbitrarily to make Q nonsingular. Let
X Px. We compute
1 0][0
1 0 1 0 1
PAP-I~[~ o1 0][1
FEo
is observable and has the same
the state equation is not controllable.

Let us choose, as in (5-56),
p- 1
and the nz-dimensional subequ
--
~,1 0~J = l~0 0:~ L~J1
Theorem 5-18 (Canonical de
Consider the linear time-invaria
FE
By equivalence transformation
10
This is a simplified version of the can

References 57.60. and 116. See also
IAMICAL EQUAnONS
e space L of FE is divided
pace of L, denoted by I: I,
I
the (n
and
C=CP-l=[l
1]
nd-dimensional
lce FEe is controllable, all
54a) shows that the state

lor indirectly through the
trollable and is dropped in
ariant dynamical equation
,tate vector can be decom
:her uncontrollable. By
obtain a controllable dy
)-state equivalent to the
[~ ~] =[1
1
0
1
203
2 : 1J
Hence, the reduced controllabfe equation is
Xc
[:
~}, +[~ ~}
y=[1
2Jx c
(5-59)
Dual to Theorem 5-16, we have the following theorem for unobservable

dynamical equations.
Theorem 5-17
Consider the n-dimensional linear time-invariant dynamical equation FE.

If the observability matrix of FE has rank n2(n2 < n), then there exists an equi
valence transformation i = Px that transforms FE into
FE:
1Jx
[x~oo_J
y
(5-58)
[Ao
0AoJ[XoJ
[DoJ
A21
Xii + Do u
[C. 0]
[:;J
(5-60a)
(5-60b)
+Eu
and the n2-dimensional subequation of FE
:[B : ADJ in determining
Xo = Aoxo +Dou
y. =Coxo+Eu
(5-61 a)
(5-61 b)
is observable and has the same transfer-function matrix as FE.
independent columns of
ake Q nonsingular. Let
~] [1 0: OJ
This theorem can be readily established by using Theorems 5-16 and 5-10.
The first n2 rows of P in Theorem 5-17 are any n2 linearly independent rows of
the observability matrix of {A, C}; the remaining n - n2 rows of P are entirely
arbitrary so long as P is nonsingular. Equation (5-60) shows that the vector
x(j does not appear directly in the output y or indirectly through xO. Hence the
vector Xi) is not observable and is dropped in the reduced equation.
Combining Theorems 5-16 and 5-17, we have the following very important
theorem.
Theorem 5-18 (Canonical decomposition theorem)
10
Consider the linear time-invariant dynamical equation
~- -~ -~-
FE:
x =Ax +Bu
y=Cx+Eu
By equivalence transformations, FE can be transformed into the following

10
This is a simplified version of the canonical decomposition theorem. For the general form, see
References 57,60, and 116. See also Reference S127.
204
CANONICAL DECOMPOSITION C
canonical form
~12 ~ An
~13][~eoJ
FE. [~eoJ_rAeo
Xeo - 0 Aeo
xeo -I-[~eoJ
' Beo U
Xc -0--- 0- , A~ - Xc
0
(5-62a)
= [0 Ceo: Cc]X +Eu
(5-62b)
-.
This theorem can be illl

which the uncontrollable p,
observable parts. We see .
response matrix of a dynan
and observable part of the
matrix (the input-output des(
controllable and observable. '
input-output description and
us why the input-output de
system, for the uncontrollabl
appear in the transfer-functic
where the vector xeo is controllable but not observable, xeo is controllable
and observable, and Xc is not controllable. Furthermore, the transfer function
of FE is
Ceo(sl -
Aeo)- 1 Beo + E
which depends solely on the controllable and observable part of the equation
FE.
Proof
If the dynamical equation FE is not controllable, it can be transformed into the
form of (5-54). Consider now the dynamical equation FEe which is the con
trollable part of FE. If FEe is not observable, then FEe can be transformed into
the form of (5-60), which can also be written as
= [AeD
[~eoJ
Xeo
0
y=
~12J[~eoJ +[~eoJ U
Aco
Xeo
Beo
[0 Cco]X + Eu
Combining these two transformations, we immediately obtain Equation (5-62).

Following directly from Theorems 5-16 and 5-17, we conclude that the transfer
function of FE is given by Ceo(sl Aco)-1Beo +E.
Q.E.D.
Example 2
Consider the network shown

source, the behaVior due to 1
detected from the output. f
L1 are not observable (they rn
the state variable associated,
metry, the state variable assocl
By dropping the state variable
the network in Figure 5-11(a)
the transfer function of the ne
Before moving to the next
of Theorem 5-7.
C1
Ll
.l
II
,-------,
l~l
co
I
I
I
I
I -
leG:
I
I
I
I
--
co
?If.
In?
In..?
J:
(a)
IJ
In
In<-
In:
L ______ --.J
Figure 5-10 Canonical decomposition of a dynamical equation. (c stands for control

lable, c for uncontrollable, 0 for observable, and 0 for unobservable.)
(b)
Figure 5-11 An uncontrollable ar
NAMICAL EQUATIONS
,] [BCD]
I
+ ~D
(5-62a)
(5-62 b)
rvable, XeD is controllable

Dore, the transfer function
205
This theorem can be illustrated symbolically as shown in Figure 5-10, in

which the uncontrollable part is further decomposed into observable and un
observable parts. We see that the transfer-function matrix or the impulse
response matrix of a dynamical equation depends solely on the controllable
and observable part of the ~quation. In other words, the impulse-response
matrix (the input-output description) describes only the part of a system that is
controllable and observable. This is the most important relation between the
input-output description and the state-variable description. This theorem tells
us why the input-output description is sometimes insufficient to describe a
system, for the uncontrollable and/or unobservable parts of the system do not
appear in the transfer-function matrix description.
vable part of the equation
Example 2
be transformed into the

ion FEe which is the con
Ee can be transformed into
Consider the network shown in Figure 5-11(a). Because the input is a current
source, the behavior due to the initial conditions in C 1 and Ll can never be
detected from the output. Hence the state variables associated with C 1 and
Ll are not observable (they may be controllable, but we don't care). Similarly
the state variable associated with L2 is not controllable. Because of the sym
metry, the state variable associated with C 2 is uncontrollable and unobservable.
By dropping the state variables that are either uncontrollable or unobservable,
the network in Figure 5-11(a) is reduced to the form in Figure 5-11(b). Hence
the transfer function of the network in Figure 5-11(a) is g(s) 1.
III
'0] U
:0
:ly obtain Equation (5-62).

conclude that the transfer
Q.E.D.
mtion. (c stands for control

ervable.)
Before moving to the next topic, we use Theorem 5-16 to prove statement 4
of Theorem 5-7.
C1
Ll
u
In
In
L2
In
In
(a)
In
In
y
In
In
(b)
Figure 5-11 An uncontrollable and unobservable system with transfer function 1.
206
Proof of statement 4 of Theorem 5-7
The matrix (sI A) is nonsingular at every sin C except at the eigenvalues of A;

hence the matrix [sI - A : B] has rank n at every s in C except possibly at the
eigenvalues of A. Now we show that if {A, B} is controllable, then
p[sI - A : B] = n at every eigenvalue of A. If not, then there exist an eigenvalue
A and a 1 x n vector ex =1= 0 such that
ex[.H
exA = exA
or
A : B] 0
and
exB =0
exA 2 = )"exA =)" 2ex
which imply
CANONICAL DECOMPOSITION 0
equation still has the same ZI

ing definition.
Definition 5-9
A linear time-invariant dym

only if there exists a linear ti
sion that has the same transi
state equivalent to FE. Otht
Theorem 5-19
A linear time-invariant dyna

i
exA = Aiex
and, in general,
is controllable and observabl
1,2, ...
Hence we have
ex[B
Proof
AB
...
An-l B]
[exB
AexB
...
An -lexB] = 0
This contradicts the controllability assumption of {A, B}. Hence, if {A, B} is

controllable, then p[sI A : B] n at every eigenvalue of A and, conse
quently, at every s in C.
N ow we show that if {A, B} is not controllable, then p[AI - A : B] < n
for some eigenvalue A of A. If {A, B} is not controllable, there exists an equi
valence transformation P that transforms {A, B} into {A, B} with
A=PAP- 1
Ac
[o
A!2]
Ac
[~,]
PB
Let A be an eigenvalue of Ac. We choose p =1=0 such that pAc = AP. Now we
form a 1 x n vector ex [0 p]. Then we have
..[ AI -A i B]
[0 II] [AI ~ A,
-A Ac Bc]
=0
0
l.:
),,1
If the dynamical equation FE

is reducible (Theorems 5-16 a
FE is controllable and obser
contradiction. Suppose tha1
and observable and that there
FE,
of lesser dimenSion, say nl < 11

Theorem 4-6, we have E = E;:
CAkB
Consider now the product
which implies
o =ex[P{AI _A)P- l : PB] =exP[(AI _A)P- 1 : BJ
VU~ ~~ 1
[
CAn-l
Since ex =1=0, we have ~~exP =1=0. Because p-l is nonsingular, ~(AI - A)P- l =0
implies a(AI A) O. Hence we have
CB
CAB
ti[AI - A : BJ = 0
In other words, if {A, B} is not controllable, then [sI A: BJ does not have a
rank of n for some eigenvalue of A. This completes the proof.
Q.E.D.
CAn-IE
By (5-64), we may replace CAk

Irreducible dynamical equation. We have seen trom Theorems 5-16 and
5-17 that if a linear time-invariant dynamical equation is either uncontrollable
or unobservable, then there exists a dynamical equation of lesser dimension
that has the same transfer-function matrix as the original dynamical equation.
In other words, if a linear time-invariant dynamical equation is either uncontrol
lable or unobservable, its dimension can be reduced such that the reduced
where Vn - 1 and Un- l are defi

lable and observable, we have
2-6 that p(VU) n. Now Vnmatrices; hence the matrix VJ
t\.MICAL EQUAnONS
207
)t at the eigenvalues of A;
C except possibly at the
I} is controllable, then
l there exist an eigenvalue
equation still has the same zero-state response. This fact motivates the follow
ing definition.
Definition 5-9
"
A linear time-invariant dynamical
equation FE is said to be reducible if and
only if there exists a linear time-invariant dynamical equation of lesser dimen
sion that has the same transfer-function matrix as FE or, equivalently, is zero
state equivalent to FE. Otherwise, the equation is said to be irreducible.
Theorem 5-19
A linear time-invariant dynamical equation FE is irreducible if and only if FE
is controllable and observable.

Proof
},n-l IXBJ
"B}. Hence, if {A, B} is

nvalue of A and, conse
then p[AI-A : BJ <n

lble, there exists an equi
{A, B} with
[!]
that
pAc=}"p.
Now we
If the dynamical equation FE is either uncontrollable or unobservable, then FE

is reducible (Theorems 5-16 and 5-17). Hence what we have to prove is that if
FE is controllable and observable, then FE is irreducible. We prove this by
contradiction. Suppose that the n-dimensional equation FE is controllable
and observable and that there exists a linear time-invariant dynamical equation
FE,
FE:
x=Ax +Bu
y=Cx +Eu
(5-63a)
(5-63b)
of lesser dimension, say n 1 < n, that is zero-state equivalent to FE. Then, from
Theorem 4-6, we have E =E and
k =0, 1,2, ...

2
~c
Boc]
~)P-l:B]
ingular, 1i(21 - A)P - 1 =0
A : BJ does not have a

he proof.
Q.E.D.
from Theorems 5-16 and

n is either uncontrollable
ltion of lesser dimension
~inal dynamical equation.
Jation is either uncontrol
d such that the reduced
(5-64)
Consider now the product
VU~ C~ 1
[
c~n-l
[B
AB
CAB
CA2B
...
An-1B]
CAn-lB
CAnB
1
(5-65)
CA2(~-1)B
By (5-64), we may replace CAkB in (5-65) by CAkB; consequently, we have
(5-66)
where Vn- l and Un- 1 are defined as in (5-48) and (5-27). Since FE is control
lable and observable, we have pU =n and pV =n. It follows from Theorem
2-6 that p(VU) n. Now Vn- 1 and Un- 1 are, respectively, qn x nl and nl x np
matrices; hence the matrix Vn-l Un- 1 has a rank of at most nl' However,
208
(5-66) implies that p(Yn-lUn-d n>nl' This is a contradiction.

FE is controllable and observable, then FE is irreducible.
CONTROLLABILITY AND oBsa
Hence, if
Q.E.D.
Recall from Section 4-4 that if a dynamical equation {A, D, C, E} has a pre
scribed transfer-function matrix G(s), then the dynamical equation {A, D, C, E}
is called a realization of G(s). Now if {A, D, C, E} is controllable and observable,
then {A, D, C, E} is called an irreducible realization of G(s), In the following we
shall show that all the irreducible realizations ofG(s) are equivalent.
Theorem 5 -20
Let the dynamical equation {A, D, C, E} be an irreducible realization of a
Then {X, B, C, E} is also an irreducible
realization ofG(s) if and only if {A, D, C, E} and {X, B, C, E} are equivalent;
that is, there exists a nonsingular constant matrix P such that X PAP - 1,
B PD, C =CP-l, and E E.
q x p proper rational matrix G(s).
invariant dynamical equatior

is referred to References 106;
*5-6 Controllabilityal
Dynamical Equations
The controllability and ob
equation are invariant unde]
ceivable that we may obtain
by transforming the equati4
equation is in a Jordan form,
almost by inspection. In th
Consider the n-dimensio
equation
Proof
The sufficiency follows directly from Theorems 4-6 and 5-15. We show now the
necessity of the theorem. Let U, V be the controllability and the observability
matrices of {A, D, c, E}, and let U, Y be similarly defined for {X, B, C, E}. If
{A, D, C, E} and {X, B, C, E} are realizations of the same G(s), then from (5-64)
and (5-65) we have E = E,
and
VU YU
VAU=YXU
(5-67)
(5-68)
where the matrices A, D, and

The n x n matrix A is in the J4
Am' Ai denotes all the Jorda
the number of Jordan blocks i
Ai = diag (Ail' Ai2' ... ,
Table 5-1 Jordan-Fonn Dynal
The irreducibility assumption implies that pY n; hence the matrix (y*y) is

nonsingular (Theorem 2-8). Consequently, from (5-67), we have
U = (y*y)-l y*VU ~ PU
(5-69)
where P~(y*y)-ly*V. From (5-69) we have pU .::;;;min (pP, pU), which,

together with pU = n, implies that pP = n. Hence P qualifies as an equivalence
transformation. The first p columns of (5-69) give B = PD. Since pU = n,
Equation (5-69) implies that
P = (UU*)(UU*) - 1
With P = (y*y)-ly*V (UU*)(UU*)-l, it is easy to derive from (5-67) and
(5-68) that V=YP and PA=XP, which imply that C=CP and X=PAP- l .
Q.E.D.
This theorem implies that all the irreducible realizations of G(s) have the
same dimension. Physically, the dimension of an irred....cible dynamical equa
tion is the minimal number of integrators (if we simulate the equation in an
analog computer) or the minimal number of energy-storage elements (if the
system is an RLC network) required\ to generate the given transfer-function
matrix.
We studied in this section only the canonical decomposition of linear time-
AI
A
(n x n) [
C [C l
Ail
(nj~in;)
C =[C il
j
(q x ni)
A.i 1
A. i
Ai}
(nij xnij)
C lj
[elij
NAMICAL EQUATIONS
COI'>l'TROLLABlLlTY AND OBSERVABILlTY OF JORDAN-FORM DYNAMICAL EQUATIONS
. contradiction. Hence, if
cible.
Q.E.D.
ion {A, B, C, E} has a pre
lical equation {A, B, C, E}
mtrollable and observable,
. G(s), In the following we
are equivalent.
reducible realization of a
E} is also an irreducible
~ B, C, E} are equivalent;
P such that A PAP-I,
*5-6 Controllability and Observability of Jordan-Form

Dynamical Equations
The controllability and observability of a linear time-invariant dynamical
equation are invariant under any equivalence transformation; hence it is con
ceivable that we may obtain simpler controllability and observability criteria
by transforming the equation into a special form. Indeed, if a dynamical
equation is in a Jordan form, the conditions are very simple and can be checked
almost by inspection, In this section, we shall derive these conditions,
Consider the n-dimensional linear time-invariant Jordan-form dynamical
equation
x=Ax +Bu
y=Cx +Eu
(5-70a)
(5-70b)
where the matrices A, B, and C are assumed of the forms shown in Table 5-l.
The n x n matrix A is in the Jordan form, with m distinct eigenvalues AI, }.z' . .. ,
Am' Ai denotes all the Jordan blocks associated with the eigenvalue }.i; r(i) is
the number of Jordan blocks in Ai; and Ai} is thejth Jordan block in Ai' Clearly,
and
(5-67)
(5-68)
lence the matrix

i7), we have
J:s;min
invariant dynamical equations. For the time-varying case, the interested reader
is referred to References 106 and 108,
JFE:
.d 5-15. We show now the

dlity and the observability
~fined for {A, B, C, E}. If
arne G(s), then from (5-64)
CV*V)
Table 5-1
Jordan-Form Dynamical Equation
is
(5-69)
[A!
A2
(n x n)
AJ
(pP, pU), which,
lualifies as an equivalence
~ B=PB. Since pU =n,
o derive from (5-67) and

C CP and A PAP- 1
Q.E.D.
iizations of G(s) have the

educible dynamical equa
mlate the equation in an
,storage elements (if the
le given transfer-function
Imposition of linear time
209
C=[C 1
Ai
(ni x n;)
Ci
~["
=[C il
C2
(n
~p) {j]
Cm]
Ai2
B.
A.J
C i2
[Bil 1
_ Bi2
(nixlp)
Bir(i)
Cir(i)]
(q x n;)
[,1,1 '.
=
Ai 1
Aij
'.
(nu x nil)
A, ~.
C ij
...
(n;jxp)
[CUi
C2ij
Bij
CliJ
[b
ll
b 2ij
.
blij
210
Let
CONTROLLABILITY AND OBSEJ
ni
and
nij
observable are that the set {

linearly independent. Alth
dent, the set {C 12I }, which
Hence J FE is not observabl
be the order of Ai and A ij, respectively; then

m
n=
rU)
I1 I1
ni
i= 1
i=
nij
j=
Corresponding to Ai and Aij' the matrices Band C are partitioned as shown.

The first row and the last row of Bij are denoted by b 1 ij and b lij , respectively.
The first column and the last column of C ij are denoted by C 1 ij and c 1ij .
Theorem 5-21
The n-dimensional linear time-invariant Jordan-form dynamical equation J FE

is controllable if and only if for each i 1, 2, ... , m, the rows of the r(i) x p
matrix
bnj
Bl~ b~:2
(5-71 a)
b,ir(i)
are linearly independent (over the field of complex numbers). JFE is observable
if and only if for each i = 1, 2, ... , m, the columns of the q x r(i) matrix
I
Ci
..
LCli1
C1i2
(5-71 b)
CUrti)
are linearly independent (over the field of complex numbers).
The conditions for COt

req uired that each of the m
pendence. The linear depe
Furthermore, the row vecto
determining the controllabil
The physical meaning of
the block diagram of the Jc
studying the general case, ,
Jordan-form dynamical eqm
an integrator and a feedback
block, or more precisely th(
variable. Each chain of bID<
Consider the last chain of F
variables in that chain can b
that chain can be observed.
same eigenvalue, then we r
vectors of these chains. The
be studied separately.
Example 1
Consider the Jordan-form dynamical equation
Al
0
0
-0 --6-:-l~: 0
X= o 0 --0-;-1-:
o 0 0 0:
o 0 0 0
o 0 0 0:
JFE:
1: 0
AI: 0
I
0
0
0
0
0
0
0
0
A2
0
0
1
A2
0
000
0
0
1 0 0 ~b111
0
o 1 0 I ~b112
0 x+IO 0 11u~b113
1 1 2
0
010
1
A2
o 0 1 ~b'21
_ _ _ .1 _ _ _ _ _ _ _ _ _
[~
1 2: 0: 0
0:1:2:0
I
1 0:2:3:0
01
2
1 1 x
(5-72a)
~
~b -~ -~
~l~
~~~
(5-72b)
2 2
iii i
C l11
C l12 C l 13 C 121
The matrix A has two distinct eIgenvalues Al and A2 There are three Jordan
blocks associated with AI; hence r(1) 3. There is only one Jordan block associ
ated with A2 ; hence r(2) = 1. The conditions for J F E to be controllable are that
the set {bill' ,12 , /13 } and the set 12 d be, individually, linearly independent.
This is the case; hence J F E is controllable. The conditions for J F E to be
b b
{b
Figure 5-12
Block diagram of
CONTROLLABILITY AND OBSERVABILITY OF JORDAN-FORM DYNAMICAL EQUATIONS
\MICAL EQUATIONS
observable are that the set {Cll I ' Cll2, C1l3} and the set {c 12 d be, individually,
linearly independent. Although the set {C I I I , C 112 ' C 113 } is linearly indepen
dent, the set {c12d, which consists of a zero vector, is linearly dependent.
Hence J FE is not observable.
then
re partitioned as shown.
)Uj and blij, respectively.
d by CI ij and CUj'
dynamical equation J F E
, the rows of the r(i) x p
(5-71 a)
bers). JFE is observable
e q x r(i) matrix
(5-71 b)
nbers).
3 0
3 0
The conditions for controllability and observability in Theorem 5-21

required that each of the m set of vectors be individually tested for linear inde
pendence. The linear dependence of one set on the other set is immaterial.
Furthermore, the row vectors of B excluding the bli/s do not play any role in
determining the controllability of the equation.
The physical meaning of the conditions of Theorem 5-21 can be seen from
the block diagram of the Jordan-form dynamical equation J FE. Instead of
studying the general case, we draw in Figure 5-12 a block diagram for the
Jordan-form dynamical equation in (5-72). Observe that each block consists of
an integrator and a feedback path, as shown in Figure 5-13. The output of each
block, or more precisely the output of each integrator, is assigned as a state
variable. Each chain of blocks corresponds to a Jordan block in the equation.
Consider the last chain of Figure 5-12. We see that if bl21 1= 0, then all state
variables in that chain can be controlled; if C l2l 1=0, then all state variables in
that chain can be observed. If there are two or more chains associated with the
same eigenvalue, then we require the linear .independence of the first gain
vectors of these chains. The chains associated with different eigenvalues can
be studied separately.
~b111
1 0
~b112
3 1
U~b113
(5-72a)
y
1 0
211
~b121
y
(5-72b)
There are three Jordan

one Jordan block associ
) be controllable are that
lly, linearly independent.
Inditions for JFE to be
Figure 5-12
Block diagram of the Jordan-form equation (5-72).
212
CONTROLLABItITY AND OBSEl
linearly independent. By p
theorem can be established.
Figure 5-13 Analog computer simulation of 1/(s -
Observe that in order f

independent, it is necessary
that is, p = I-it is necessar
linearly independent. In wo
form dynamical equation t4
block associated with each d
to a vector. Thus we have 1
AJ
Proof of Theorem 5-21

We use statement 4 of Theorem 5-7 to prove the theorem. In order not to be
overwhelmed by notation, we assume [sI - A : B] to be of the form
s - A1
-1
-A 1
-1
S
A1
b ll1
b 211
o
,
----------------------:
- --i: -----. .:
0
bl21
A single-input linear time-i:

troll able if and only if ther
distinct eigenvalue and all tl
spond to the last row of eac
A single-output linear tl
observable if and only if th(
distinct eigenvalue and all th
to the first column of each J,
(5-73)
Example 2
bl1 1
b 11 2
j - - -:
:
s A1 :
, b'12
- - - - - - - - - - - - - - - -: -s- -)~ - - - - - -=..-( - -,- -bl-2~
S
A2
The matrix A has two distinct eigenvalues A1 and A2 . There are two Jordan
blocks associated with Ab one associated with A2 If SAl, (5-73) becomes
-1
0
0
0
0
0
- - - - - - - - - - -
0
-1
0 ,,
- - - - T - - ,0
'0
Corollary 5-21
Consider the single-variable
b ll1
X=
b 211
bill
- - - - - - - -,
-1,
b l12
0,
(5-74 )
l12
-----------T----------------T-----
, A -A
-1, b
A1
121
A2 ,: b /21
By a sequence of elementary column operations, the matrix in (5-74) can be

transformed into
1 0:
I
0 1I
~--~--~H
There are two distinct eigen

responds to the last row of
is zero; therefore, the equati(
corresponding to the first coli
therefore, the equation is obs
Example 3
o
o
-1
Consider the following two J
- - - - - - - - - - - - - - i -
------------:- i: =--i
: 0
:--0--
[ ~1-
bill
0- --- =--{ -:
- 2
- - -
(5-75)
b ll2
0- ---:- -0--
,
A1 - A2 : 0
Note that A1 -A 2 is different from zero. The matrix in (5-75), or equivalently,

the matrix in (5-73) at S = Ab has a full row rank if and only if bIll and b l12 are
X2_
and
~1X2_
That the state equation (5<

Equation (5-77) is a time-va
matrix is in the Jordan form
MICAL EQUA nONS
CONTROLLABIUTY AND OBSERVABILITY OF JORDAN-FORM DYNAMICAL EQ\JATlONS
213
linearly independent. By proceeding similarly for each distinct eigenvalue, the

theorem can be established.
Q.E.D.
Observe that in order fort the rows of an r(i) x p matrix B~ to be linearly
independent, it is necessary that r(i)sp. Hence in the case of single inputthat is, p = I-it is necessary to have r(i) = 1 in order for the rows of B~ to be
linearly independent. In words, a necessary condition for a single-input Jordan
form dynamical equation to be controllable is that there is only one Jordan
block associated with each distinct eigenvalue. For p = 1, the matrix B~ reduces
to a vector. Thus we have the following corollary.
em. In order not to be
be of the form
b lll
b 211
bill
b 112
-------------~-~~~
-1: b
-A 2
l2l
s -A2 : b /21
(5-73)
There are two Jordan
Corollary 5-21
A single-input linear time-invariant Jordan-form dynamical equation is con
trollable if and only if there is only one Jordan block associated with each
distinct eigenvalue and all the components of the column vector B that corre
spond to the last row of each Jordan block are different from zero.
A single-output linear time-invariant Jordan-form dynamical equation is
observable if and only if there is only one Jordan block associated with each
distinct eigenvalue and all the components of the row vector C that correspond
to the first column of each Jordan block are different from zero.
Example 2
Consider the single-variable Jordan-form dynamical equation
; =Ab (5-73) becomes

b lll
b 2ll
X=
bill
b 112
I
(5-74)
b ll2
--T-----
'2
b l2l
b'21
matrix in (5-74) can be
OO:I]x
Consider the following two Jordan-form state equations:

(5-75)
: b ll2
---------
: 0
I
1 (5-75),
y=[1
There are two distinct eigenvalues 0 and 1. The component of B which cor
responds to the last row of the Jordan block associated with eigenvalue 0
is zero; therefore, the equation is not controllable. The two components of C
corresponding to the first column of both Jordan blocks are different from zero;
therefore, the equation is observable.
Example 3
o
o
'2
~[ ~ ~ i~]. + [1+
~l
~ __~ __~_!~
or equivalently,
only if bill and b ll2 are
[~:J=[ -~ -~Jx +[:}
[xXlJ = [-10 - OJ2 x +[e-e - J

u
(5-76)
and
2t
(5-77)
That the state equation (5-76) is controllable follows from Corollary 5-21.
Equation (5-77) is a time-varying dynamical equation; however, since its A
matrix is in the Jordan form and since the components of B are different from
214
OUTPUT CONTROLl
where A, B, and Care n x n

response matrix of the systel
zero for all t, one might be tempted to conclude that (5-1'1) is controllable.
Let u~ .;heck this by using Theorem 5-4. For any fixed to, we have
0
-(to-t)
c)(t o
t)B(t)= [ eO
e-
2o
(t -t)
J[:-2t
-t ]
[-to]
:-2t
The transfer-function matrix

It is clear that the rows of c)(t o - t)B(t) are linearly dependent in t. Hence the
state equation (5-77) is not controllable at any to.
From this example we see that, in applying a theorem, all the conditions
should be carefully checked; otherwise, we might obtain an erroneous con
clusion.
It is clear that G(s) is a strict
*5-7 Output Controllability and Output Function

Controllability
A system whose transfer fun<

output controllable if and OJ
over the field of complex nUll
Similar to the (state) controllability of a dynamical equation, we may define

controllability for the output vector of a system. Although these two concepts
are the same except that one is defined for the state and the other for the output,
the state controllability is a property of a dynamical equation, whereas the
output controllability is a property of the impulse-response matrix of a system.
Consider a system with the lilput-output description
y(t)
~ G(t,r)u(r) dr
where u is the p x 1 input vector, y is the q x 1 output vector, G(t, T) is the q x p

impulse-response matrix of the system. We assume for simplicity that G(t, T)
does not contain t>-functions and is continuous in t and T for t > T.
Corollary 5-22
[CD
has rank q.
The proof of Theorem 5-~

trivial consequence of this co
controllable. We see that
can also be stated in terms a
independence of G(s), the COil
The state controllability
the output controllability is c
these two concepts are not nc
4
Definition 5-10
A system with a continuous impulse-response matrix G(t, T) is said to be output

controllable at time to if, for any Yl' there exist a finite t 1 > to and an input
that transfers the output from y(t o) = 0 to y(t d = Y1
u[to,td
Example 1
Consider the network shown

observable, but it is output a
Theorem 5-22
A system with a continuous G(t, T) is output controllable at to if and only if there

exists a finite t 1 > to such that all rows of G(t 1, T) are linearly independent in r
on [to, t 1] over the field of complex number.
The proof of this theorem is exactly the same as the one of Theorem 5-4
and is therefore omitted.
We study in the following the class of systems tha~ also have linear time
invariant dynamical-equation descriptions. Consider the system that is
describable by
FE:
x=Ax +Bu
y=Cx
tn
y
tn
Figure 5-14 A network which i

observable.
OUTPUT CONTROLLABILITY AND OUTPUT FUNCTION CONTROLLABILITY
MICAL EQUATIONS
Lt (5-77) is controllable.
1 to, we have
-to]
[:-2t
215
where A, B, and Care n x n, n x p, q x n real constant matrices. The impulse

response matrix of the system is
The transfer-function matrix of the system is
)endent in t. Hence the
orem, all the conditions

,tain an erroneous con-
(5-78)
It is clear that G(s) is a strictly proper rational function matrix.
Corollary 5-22
A system whose transfer function is a strictly proper rational-function matrix is
output controllable if and only if all the rows of G(s) are linearly independent
Function
over the field of complex numbers or if and only if the q x np matrix
:quation, we may define

ough these two concepts
l the other for the output,
1 equation, whereas the
lonse matrix of a system.
In
{ector, G(t, r) is the q x p

or simplicity that G(t, r)
drfort>r.
[CB: CAB: ... : CAn-1BJ
(5-79)
has rank q.
The proof of Theorem 5-7 can be applied here with slight modification. A
trivial consequence of this corollary is that every single-output system is output
controllable. We see that although the condition of output controllability
can also be stated in terms of A, B, and C, compared with checking the linear
independence of G(s), the condition (5-79) seems more complicated.
The state controllability is defined for the dynamical equation, whereas
the output controllability is defined' for the input-output description; therefore,
these two concepts are not necessarily related.
;(t, r) is said to be output
Example 1
lite t 1 > to and an input
Consider the network shown in Figure 5-14. It is neither state controllable nor
observable, but it is output controllable.
=Yl'
le at to if and only if there

linearly independent in r
the one of Theorem 5-4
at also have linear time

ier the system that is
In
u
1+
I
In
L - -_ _ _ _ _- - - - - - '
Figure 5-14 A network which is output controllable but neither (state) controllable nor
observable.
216
CONTROLLABILITY AND OBSERVABILITY OF LINEAR DYl'T:\MICAL EQUATIONS
range of G(s), for which th

-----1
. YI
~Y2
Figure 5-15 A system which is controllable and observable but not output controllable.
Example 2
Consider the system shown in Figure 5-15. The transfer-function matrix of

the system is
s
[
;1].
G(s)G.
s +1
the rows of which are linearly dependent. Hence the system is not output
controllable. The dynamical equation of the system is
If the input is restricted

then the given output func
computed from (5-81) will]
given output has some diSCI
needed to generate the disc(
The condition for outp
terms of the matrices A, B, a
interested reader is referred
Dual to the output functil
These problems are intimat,
q x p proper rational matri:
exists a p x q rational matriJ
-x +u
Y~G}
which is controllable and observable.
If a system is output controllable, its output can be transferred to any

desired value at certain instant of time. A related problem is whether it is
possible to steer the output following a preassigned curve over any interval of
time. A system whose output can be steered over any interval of time is said
to be output function controllable or functional reproducible.
Theorem 5-23
A system with a q x p proper rational-function matrix G(s) is output function

controllable if and only if pG(s) = q in ~(s), the field of rational functions with
real coefficients.
Proof
If the system is initially relaxed, then we have

y(s) = G(s)u(s)
(5-80)
If pG(s) q-that is, all the rows of G(s) are linearly independent over the
field of rational functions-then the q x q matrix (;{s)G*(s) is nonsingular
(Theorem 2-8). Consequently, for any y(s), if we ,choose
u(s) = G*(s)(G(s)G*(S-l y(s)
(5-81)
then Equation (5-80) is satisfied. Consequently, if pG{s) = q, then the system is

output function controllable. If pG(s) < q, we can always find a y(s), not in the
A system is said to have an ir

A necessary and sufficient (
pG(s) q in [~(s). This cor
controllability. Many quest
unique? Is it a proper rati
stable? What are its equiv
problems will not be stuQie(
References S172, S185, S218,
5-8 Computational I
In this section, we discuss s(

chapter. As discussed in Sec
conditioned; a computationa
unstable. If we use a nume
problem, the result will gener
if we use a numerically stable
res ult will be correct. If WI
problem, well or ill condition
a problem, if we must use an
stable method, the unstable
possible, in the computation.
As discussed in Theorem
trollability of a state equatioJ
more suitable for computer a
computational problems, as,
The computation of the c
is straightforward. Let Ko ~
COMPUTATIONAL PROBLEMS
IAMICAL EQUATIONS
217
range of (;(s), for which there exists no solution o(s) in (5-80) (Theorem 2-4).
Q.E.D.
Ie but not output controllable.
:ansfer-function matrix of
the system is not output

is
:an be transferred to any

problem is whether it is
curve over any interval of
.ny interval of time is said
'ucible.
'ix (;(s) is output function

of rational functions with
(5--80)
.fly independent over the

(;(s )(;*(s) is nonsingular
ose
(5--81 )
~(s) = q, then the system is
ways find a y(s), not in the
If the input is restricted#, to the class of piecewise continuous functions of t,

then the given output function should be very smooth; otherwise, the input
computed from (5-81) will not be piecewise continuous. For example, if the
given output has some discontinuity, an input containing <5-functions may be
needed to generate the discontinuity.
The condition for output function controllability can also be stated in
terms of the matrices A, B, and C. However, it is much more complicated. The
interested reader is referred to Reference 8.
Dual to the output function controllability is the input function observability.
These problems are intimately related to the inverse problem. A system with a
q x p proper rational matrix (;(s) is said to have a right (left) inverse if there
exists a p x q rational matrix GR1(S) [Gu(s)] such that
A system is said to have an inverse if it has both a right inverse and a left inverse.
A necessary and sufficient condition for (;(s) to have a right inverse is that
p(;(s) = q in !R(s). This condition is identical to that of the output function
controllability. Many questions may be raised regarding a right inverse. Is it
unique? Is it a proper rational matrix? What is its minimal degree? Is it
stable? What are its equivalent conditions in dynamical equations? These
problems will not be studied in this text. The interested reader is referred to
References S172, S185, S218, and S239.
* 5-8
Computational Problems
In this section, we discuss some computational problems encountered in this

chapter. As discussed in Section 2-9, a problem may be well conditioned or ill
conditioned; a computational method may be numerically stable or numerically
unstable. If we use a numerically stable method to solve a well-conditioned
problem, the result will generally be good. If a problem is ill conditioned, even
if we use a numerically stable method to solve it, there is no guarantee that the
result will be correct. If we use a numerically unstable method to solve a
problem, well or ill conditioned, the result must be carefully scrutinized. Given
a problem, if we must use an unstable method because of nonexistence of any
stable method, the unstable method should be applied at a stage, as late as
possible, in the computation.
As discussed in Theorem 5-7, there are several ways of checking the con
trollability of a state equation. Among them, statements 3 and 4 appear to be
more suitable for computer computation. However, they may encounter some
computational problems, as will be discussed in the following.
The computation of the controllability matrix U = [B AB ... An-IB]
is straightforward. Let Ko~B. We compute Ki =AK i - l , i = 1,2, ... , n -1.
218
At the end, we have U = [Ko Kl

Kn - l ]. The rank of U can then be
computed by using the singular value decomposition (Appendix E) which is a
numerically stable method. If the dimension n of the equation is large, this
process requires the computation of AkB for large k and may transform the
problem into a less well ,conditioned problem. For convenience in discussion,
we assume that all eigenvalues Ai of A are distinct and B is an n x 1 vector.
We also arrange Ai so that 1..1,11 ;;::::IA21 ; : : ... ;;::::IAnl. Clearly, we can write B as
F or example, consider the r
A=
B=(XI VI + (X2V2 + ... + (Xnvn
where Vi is an eigenvector associated with eigenvalue Ai; that is, AVi =

straightforward to verify
AiVi,
It is
AkB =(Xl"~.1Vl +(X2A~v2 + ... +(XnA~n
(see Reference S212). The di

on the superdiagonal are all
except the e at the (20, l)th pc
computed as
If 1..1,11 is much larger than all other eigenvalues, then we have

AkB-+(XIA~Vl
.1{
for k large
In other words, AkB tends to approach the same vector, VI' as k increases.
Hence, it will be difficult to check the rank of U if n is large.
The same conclusion can also be reached by using a different argument. The
condition number of a matrix A may be defined as cond A~ IIAI1211A -1112 = (It/(ls,
where (I, and (Is are the largest and smallest singular values of A. It can be shown
that (ls~IAnl<IAll~(ll' Hence, if IAll~IAnl, Cond A is a very large number.
In computer computation, the multiplication of a matrix with a large condition
number will introduce large computational error and should be avoided.
Hence, the use of [B AB ... An-1B] to check the controllability of {A, B}
is not necessarily a good method.
As an example consider the 10-dimensional state equation (see Reference
S169)
*=r
lo
2-
If e 0, the eigenvalues of A
the migration of the eigenval
5-16, the root locus of .1(s) (sc
indicate the migration of the j
eigenvalues except Al and A20
1J
This analysis was suggested by Pn
0] +[:]u
We compute U = [B AB ... A B] and then compute its singular values.

The three smallest singular values are 0.712 x 10- 7 , 0.364 X 10- 9 , and
0.613 x 10- 12 . If we use a digital computer with a relative preCision of 10- 10
(or a number will be considered as a zero if its absolute value is smaller than
10- 10 ), then the rank of U is smaller than 10, and we will conclude that the
equation is not controllable, although the equation is clearly controllable
f
following Corollary 5-21.
If we use the condition rank [sl - A BJ n, for all eigenvalues of A, to
check the controllability of {A, B}, we must compute first the eigenvalues of A.
The QR method is a stable and reliable method of computing the eigenvalues of
a matrix. However, the eigenvalues of some matrices can be very ill conditioned.
I
Figure 5-16
The foot locus of
~AMICAL
EQUAnONS
he rank of U can then be

I (Appendix E) which is a
:he equation is large, this
k and may transform the
::onvenience in discussion,
and B is an n x 1 vector.
early, we can write B as
COMPUTATlONAL PROBLEMS
219
For example, consider the matrix

20
20
19
A=
20
18
0
20
20
1
(see Reference S212). The diagonal elements range from 1 to 20; the elements
on the superdiagonal are all equal to 20. The rest of the matrix are all zeros
except the 8 at the (20, 1)th position. The characteristic polynomial of A can be
computed as
n
20
'He have
~(s) =
(s
i) - (20)19 8
i= 1
vector, VI' as k increases.
is large.
,a different argument. The
dA~IIAI121IA -1112 = (Jd(Js'

tlues of A. It can be shown
A is a very large number.
Ltrix with a large condition
and should be avoided.
Ie controllability of {A, B}
If 8 0, the eigenvalues of A are clearly equal to Ai = i, i = 1,2, ... ,20. To see

the migration of the eigenvalues of A as 8 increases from 0, we plot, in Figure
5-16, the root locus of ~(s) (see Reference S46).11 The heavy lines with arrows
indicate the migration of the eigenvalues as 8 increases from O. We see that all
eigenvalues except Al and A20 become complex if 8 is suffiCiently large (actually
II
This analysis was suggested by Professor D. Z. Zheng.
te equation (see Reference
Impute its singular values.

10- 7 , 0.364 x 10- 9 , and
relative precision of 10- 10
olute value is smaller than
we will conclude that the
ion is clearly controllable
:or all eigenvalues of A, to
e first the eigenvalues of A.
Imputing the eigenvalues of
; can be very ill conditioned.
Figure 5-16
The root locus of ~(s).
220
if e is larger than 7.8 x 10- 14). We list in the following A1 0 and All for some e:
= 7.8 x 10- 14
e2 = 10- 10
el
10- 5
e4= 1
e3 =
10.5 jO
10.5 j2.73
10.5 j8.05
10.5 j16.26
P1B
x x
o x x
10 0 x
:1=[Bn
(5-82)
A(1) : A(1)J
(5-83)
PI AP'l = [ "A<lA -:- A[11
21! 22
where A\11 is an nl x nl matrix, A~11 is an (n - nd x nl matrix, and so forth.
If AW 0, the controllability matrix
B(I) A(1)B(1) (A(1)fB(l) ., 'J
P 1 [ __1 - - - - ! ~ - ~ - - - - ! ~ - - ! - - - - o
0
0
...
(5-84)
Inl
OJ P
P2
where P 2A~ldP; has been pa

if A~2d = 0, the controllabilit:
U = [Inl
0
P2
has rank nl + n2 (where x de

lable and can be reduced,
is transformed, by a sequencl
A~ PAP' =1
L
where B\1) is an nl x P upper triangular matrix and n1 = rank B. This step can
be accomplished by a sequence of Householder transformations. 13 We then
compute
U = [B AB:: An - 1 B]
and
:I
000
If AW =foO, we
where AW is a n2 x nl mat
They are computed directly from A by using the QR method. [The computation
of the eigenvalues from 1\(s) is not advisable because the characteristic poly
nomial may be more sensitive to e than the eigenvalues and the roots of a poly
nomial are very sensitive to the variations of the coefficient.] We see that the
eigenvalues are very sensitive to e. Thus the eigenvalues of A are very ill
conditioned. For this reason, the use of the criterion rank [sl A B] = n
to check the controllability of {A, B} may yield an erroneous result.
It turns out that the use of Theorem 5-16 is the best way of checking the
controllability of a state equation. Furthermore, if the equation is not control
lable, the computation also yields a reduced controllable equation. The proof
of Theorem 5-16 provided a procedure of computing the required equivalence
transformation. The procedure, however, uses the controllability matrix and is
not satisfactory from the computational point of view. In the following, we
shall introduce an efficient and numerically st~ble method to transform a state
equation into the form in (5-54). For the single-input case, the method is
essentially the procedure of transforming the matrix A into the Hessenberg
form (see Appendix A). We discuss in the following the general case.
Let PI be an orthogonal matrix,12 that is, Pi I = P't. such that
x
{A\11, B\l)}.
All
A21
0
A22
A32
C=CP' =[C I
A12
C2 C
where pB\I) = nf, pA21 = n2, ,

smallest integer such that ei
Ak,k - I has dimension nk x nk.
p ?:.,
IfnI +n2 +, .. +nk n, {A, 1

By this process, the controllal
The A in (5-89) is in the bic
formation of {A, B} into the
follows (see Reference S203):
Step 1. P 1m Ao =A, Bo =E
Step 2. Find an orthogonal tr
'I
has rank n 1 < nand {A, B} is not controllable and can be reduced to controllable
{A, B} has complex elements, we use a unitary matrix, that is, PIt P'f.
13The PI can also be obtained by using gaussian elimination with partial pivoting. In this case,
P t is not orthogonal. However, the method is still numerically stable.
12lf
where rank
Zj=nj.
14 The form of A is identical to the om
IAMICAL EQUATIONS
19 A10 and All for some B:
{A~ll, B~l)}.
If A~ll +0, we find an orthogonal matrix P 2 such that
~ [ A~21J
p 2A W
3
5
26
where A~zl is a n2 x n1 matnx with nz = rank
AW and
n2:::::; n1.
(5-88)
has rank n1 + n2 (where x denotes nonzero matrices), and {A, B} is not control
lable and can be reduced. If A~2(1=0, we continue the process until {A, B}
is transformed, by a sequence of orthogonal transformations, into 14
All
A21
A~PAP'= 0
Alk
B\l)
AZk
A33
A 1 ,k-1
AZ,k-1
A 3,k-1
A3k ,B~PB=
0
0
Ak,k-l
Ak,k
A12
A22
A32
A13
A 23
C=CP' =[C 1 C 2 C 3
...
C k-
nl matrix, and so forth.
~lV~~!l~
o - ~.~.~J.
(5-84)
where pB\l) =n1, pA21 n2, pA32 = n3,"" pAk,k-1 nk' The integer k is the
smallest integer such that either nk 0 or nl +n2 +... +nk =n. Note that
Ak,k-l has dimension nk x nk-1 and has a full row rank. Clearly, we have
p;:::nl;:::nZ;:::n3'"
;:::nk;:::O
+ ... +nk n, {A, B} is controllable; otherwise, it is not controllable.

By this process, the controllability of {A, B} can be determined.
The A in (5-89) is in the block Hessenbergform (see Appendix A). The trans
formation of {A, B} into the form in (5-89) can be carried out recursively as
follows (see Reference S203):
Step 1. PIn' Ao A, Bo = B, Ii = 0, j = 1.
Step 2. Find an orthogonal transformation P j such that
be reduced to controllable
'. Pi
where rank Z j
nj. If nj = 0, go to step 7.
PT.
h partial pivoting. In this case,
stable.
(5-89)
C k]
If n1 +n2
I{
(5-86)
Now
(5-82)
(5-83)
We compute
(5-87)
1 = rank B. This step can

nsformations. 13 We then
(5-85)
:J P1B~[~\1]
[I~,
~thod.
[The computation
~e the characteristic poly
es and the roots of a poly
fficient.] We see that the
nvalues of A are very ill
ion rank [sI -A B] n
Toneous result.
best way of checking the
le equation is not control
able equation. The proof
~ the required equivalence
,ntrollability matrix and is
ew. In the following, we
ethod to transform a state
nput case, the method is
lx A into the Hessenberg
the general case.
P'l, such that
221
14
The form of A is identical to the one in (A-17).
222
Step 3. Compute
P }}A . 1 p'.}
y.]
X.
- J - - Aj
- "
[B
j
where Xj is a nj x nj matrix.
Step 4. Update
_
P-P 0
[In
0]
p.
}
Step 5. n=n + nj. If n n, go to step 8.
Step 6. j
+ 1 and go to step 2.
Step 7. {A, B} is not controllable and can be reduced to an n-dimensional
controllable equation.
Step 8. {A, B} is controllable.
This is a numerically stable method of checking the controllability of a state
equation. This algorithm was first proposed by Rosenbrock [S185] and then
discussed in References S6 and S155. These papers used the gaussian elimina
tion with partial pivoting in finding Pj' Van Dooren [S203] and Patel [S170]
suggested to use orthogonal transformations to improve the numerical stability
of the algorithm. The singular value decomposition was also suggested in
Reference S170 to determine the rank of Bi . .
The algorithm actually reveals more than the controllability of {A, B}; it
also reveals its controllability indices. A comparison of (5-84) and (5-88) with
(5-28) yields immediately
; =0,1,2, ...
ri=p-ni+l
w;here rj is the number of linearly dependent columns in AiB and ni+ 1 is the
number of linearly independent columns in AiB. Since the set of controllability
indices is uniquely determinable from {ri' i 0, 1, 2, ... } [see Figure 5-7 or
Equation (5-33)], it is also uniquely determinable from {ni' i = 1,2, ... }.
Actually we can say a little more about the controllability indices. In order
not to be overwhelmed by notation, we assume that {A, B}, where A and B
are respectively 8 x 8 and 8 x 3 matrices, has been transformed into
:J)
o :1:
x
"
o 0 L'l\
------000
000
000
000
000
X X
1,- - - -
:J)
x:
x:
,'l\.:
x:
o
o
B= 10--6--6
000
000
000
_0--6--6
The matrix A is said to be if
This is achieved by c1
form.
1
1
L _ _
x x
1 x
0 1
p- 1
x'x x'x x'x

x:x x:x
x:x
,
x'x
x'x
x'x
.J _ _ _ _ _
- -
controllability indices can a

computation. The second
dependent on its left-hand
dependent; hence 112 = 1. I
not appear in the subsequer
of A, both columns are line
A; hence A2 b1 and A2 b3 are
A, the first column is lineal
A3 b1 is linearly dependent,
column is linearly indepen4
linearly independent of its I
these indices are equal to t
triangles as shown in (5-90).
The {A, B} in (5-90) can
tion into the follOWing form
1
1
1
x:
x:-:-
0
X
X
X
- - - - --- -- ~ ,;.;', - - - ~ - - - - -
o 0 0 ,\ Lx, X X X
o 0 0: j\: hI c 1 : d 1
(5-90)
Then we have
(f - 0-
-6 -:-6 - 0-,: 0- =/E:fl
'--v----'
1st
'-v--'
'-v--'
2nd
3rd
block column
Let B = rb l b2 b3 1; that is, bi is the ith column of B. Then from the structure
of A and B in (5-90), we can readily verify that III 3,112 1, and 113 =4. These
B1 =P l B B
Al PIAl
iAMICAL EQUATIONS
223
red to an ii-dimensional
le controllability of a state
;enbrock [S185J and then
lsed the gaussian elimina
1 [S203J and Patel [S170J
we the numerical stability
m was also suggested in
controllability indices can actually be read out directly from {A, B} without any
computation. The second column of the first block column of A is linearly
dependent on its left-hand-side columns. This implies that Ab 2 is linearly
dependent; hence /12 = 1. Once a column becomes linearly dependent, it will
t
not appear in the subsequent block columns of A. In the second block column
of A, both columns are linearly independent of their left-hand-side columns of
A; hence A2 b1 and A2 b3 are linearly independent. In the third block column of
A, the first column is linearly dependent on its left-hand-side columns; hence
A3 b1 is linearly dependent, and /11 3. The second column of the third block
column is linearly independent, and there is no other column of A which is
linearly independent of its left-hand-side columns; hence /13 4. We see that
these indices are equal to the numbers of 1 enclosed by circles, squares, and
triangles as shown in (5-90).
The {A, B} in (5-90) can be further transformed by equivalence transforma
tion into the following form:
x,x
x:x x,x
,
x x x:x x:x x;x
x:x x, X x, x
X
X
Introllability of {A, B}; it

1 of (5-84) and (5-88) with
0 1 x
0 0
------0 0 0
0 0 0
------000
0 0 0
-----~-----~----~--
0 0:0 0: 0 0:0
0
0 0 1 :0 0:0 __0:.J __
-------"'-----1..
0 0 0: 1 0;0 0: 0
0:J __0
.J _ _ _ _1
_ '- 0
____
0 0 __0:0
0--6 0 ;0 0:0 1 0
A=
(5-91 )
I
I
o -6--6
IS in AiB and ni+ 1 is the

e the set of controllability
, ... } [see Figure 5-7 or
I
I
The matrix A is said to be in the block companion form or the block Frobenius
form. This is achieved by choosing
n{ni,i=1,2, ... }.
lIability indices. In order

t {A, B}, where A and B
nsformed into
,
P 1- l _
-
:x
:x
x:x
x:x
P1=
'x x'x
_ L ___ __ 1_ _
:x x:x
I
'x
I
I
X 'x
~: b
;0
d1
/l\~ :f1
i C1:
!.. ___
A _ '.
(5-90)
_ L ____ _ 1_ _
fl
1
'-.r-'
3rd
o
Then we have
x'x x'x
x:x x:x x:x

XiX
x'x
x:x
!.. _____
x:x x:x x:x
l;x x;x x'x
x X
1 x
00
-
_I _ _ _ ... _
1_ _ _
Imn
0--6--6:-1-~-:~--~-:-x
Then from the structure

112 = I, and /13 = 4. These
o 0 0:0 1 :b 2
C2
;d 2
o -6--6 ~ -0- -If :*0- i -:"0
(5-92)
224
CONTROLLABILITY AND OBSERV ABILITY OF LINEAR DYNAMICAL EQU\ III '"''
Proceeding upward, {A, B} in (5-90) can be transformed into the form in (5-91).
This process can be easily programmed on a digital computer (see Reference S6).
The process of transforming {A, B} into the block Hessenberg form in (5-89)
is numerically stable. The process of transforming the block Hessenberg form
into the block companion form in (5-91), however, is not numerically stable.
The matrix Pi 1 in (5-92) carries out gaussian elimination on columns of A
without any pivoting. If pivoting is used, the form of A will be altered, and we
can never obtain the form in (5-91). Hence, Pi 1 must be chosen without any
pivoting and the process is numerically unstable.
The {A, A} in (5-90) or, equivalently, the {A, B} in (5-91) has the con...trq.1
lability indices 111 = 3, 112 1, and 113 4. Based on these indices, the {A, B}
can be transformed into
1 x
0 0
0 0
x,x,x
x x
, ,
0
0
0 0:0:0 0 0 0
1 0:0:0
0 0 0
0
_______
__________
X
X
x:x:x x x x
-------'--r----------I
x X x,x'x
x x X
, ,
0
0 0
0 0 o:0 : 1
0 0 0:0:0
1 0 0
, ,
0
1 0
0 0 0:0:0
00'
~~_L
-------
B = PH = I?--~ - -~ I
PAP- 1
000
000
000
are left coprime. Similar t

-A
I
T"= [
B:
0:
0:, -A B:,
~ ~ ,~ ~ ~ ,~
0 :
0:
'-v---'
________
ro
r1
There are 11 + 1 block colur

A and I and p columns form
B columns. Now we searcl
left to right. Because of tl
independent of their left-ha
dependent B columns in the
of the structure of TIL' we ha
(5-93)
O~
Let 11 be the least integer SUt

0<
and
by the equivalence transformation 15
(0 0 0 0 0 0 0 0
o
P
0 0
o 0
o ~1~
o
o
o
o
L.. -...
0
0
0
0
0
0
c:D
0 0
0 0 ':(: 0
0 0 0 0
0
0
rank V = rank [B AB .
= total number of 1:
1\
:!~ 0
0
0 0 /!\
0 0 0
0 0 0
0 0 0
0 0 0
0 :i\ 0
0 0 :i~
The positions of 1 in this matrix are determined from the positions of 1 in

(5-90) with the same encirclements. The transformation of {A, H} into (A, B}
can also be easily programmed without first determining the controllability
indices (see Reference S6). The form in (5-93) is said to be in the controllable
form and is very useful in the design of state feedback as will be seen in Chapter 7.
We mention one more method of checking controllability to conclude this
section. From statement 4 of Theorem 5-7 and Theorqn G-8', we may conclude
that {A, B} is controllable if and only if the polynomial matrices sI - A and B
15
It can be shown directly (Pn

in (5-94) are the same as the
This performs only permutations of columns and rows.
Consequently, we conclude'
number of linearly independ(
independent columns of TIL a
position of the rows of TIL' h
apply Householder transfol
pivoting on the rows of TIL tc
Appendix A). Once in this.
readily determined.
There are two disadvant
{A, B}. First, the size ofT/l i
larger than the size of A. ~
controllable equation cannot
of {A, B} into a block Hessen1
the controllability of {A, B}.
The discussion of the ot
part and will not be repeated.
AMICAL EQU\ I II I"'"
ed into the form in (5-91).

mputer (see Reference S6).
Hessenberg form in (5-89)
le block Hessenberg form
IS not numerically stable.
lnation on columns of A
Awill be altered, and we
st be chosen without any
in (5-91) has the control

these indices, the {X, B}
):0 0 0 0
)_ L: __
0 0 0 0
(:x x x X
--
-r-----
x X X X
): 1 0 0 0
):0
0 0
0
):0 0
225
are left coprime. Similar to Theorem G-14, we form the matrix
-A B:
Til
0:
0: -A B:
0: I 0:1.
I
..
..
I
I
..
I
..,
..
..
..
0 :
0:
i-~ ~l
(5-94)
(no. of dependent columns)
There are /l + 1 block columns in Til; each consists of n columns formed from
A and I and P columns formed from B. We call the former A columns, the latter
B columns. Now we search linearly independent columns of Til in order from
left to right. Because of the unit matrix I, all A columns in Til are linearly
independent of their left-hand-side columns. Let fj be the number of linearly
dependent B columns in the (i + 1)th block column as shown in (5-94). Because
of the structure of Til' we have
(5-93)
Let /l be the least integer such that
and
o
o
o
o
o
o
'~
"
,'I'..
'om the positions of 1 in
tion of {A, B} into (A, B}
nining the controllability
I to be in the controllable
.s will be seen in Chapter 7.
ollability to conclude this
em G-8', we may conclude
lial matrices sI - A and B
OSfOSfl S'" Sfll-I <P

fll
fll+1="'=P
It can be shown directly (Problem 5-35) or deduced from Chapter 6 that the fi
in (5-94) are the same as the fj in (5-:28). Hence we have
rank U=rank [B AB ... All-IB] =(p fo) +(p ftl + ... +(p
= total number of linearly independent B columns in (5-94)
ll
Consequently, we conclude that {A, B} is controllable if and only if the total

number of linearly independent B columns in (5-94) is n. Note that the linearly
independent columns of Til are to be searched in order from left to right. The
position of the rows of Til' however, can be arbitrary altered. Hence, we may
apply Householder transformations or gaussian eliminations with partial
pivoting on the rows of Til to transform Til into an upper triangular form (see
Appendix A). Once in this form, the linearly independent B columns can be
readily determined.
There are two disadvantages in using Til to check the controllability of
{A, B}. First, the size of Til is (2 +/l)n x (n +p)(/l +1), which is generally much
larger than the size of A. Second, if {A, B} is not controllable, its reduced
controllable equation cannot be readily obtained. Hence, the transformation
of {A, B} into a block Hessenberg form seems to be a better method of checking
the controllability of {A, B}.
The discussion of the observability part is similar to the controllability
part and will not be repeated.
226
5-9
Concluding Remarks
Similar to Definition 5-1, ,

given any Xo and any Xl> t1
transfer Xo to Xl' 16 The c
An-IB] n (see Problem
Theorem 5-7. Hence, mos
case are applicable to the
the time-varying case, the!
discussed.
In this chapter we have introduced the concepts of controllability and observa

bility. Various theorems for linear dynamical equations to be controllable and
observable were derived. We shall discuss briefly the relations among some of
these theorems. We list first those t'leorems which are dual to each other:
Controllability:
Theorems
Observability:
Theorems
TT T T T T
6
5-9
5-11
5-12
5-13
5-14
f~~orem
5-17
The theorems in the observability part can be easily derived from the con
trollability part by applying Theorem 5-10 (theorem of duality), and vice versa.
Theorems 5-1 and 5-4 (or 5-9) are two fundamental results of this chapter.
They are derived with the least assumption (continuity), and hence they are
most widely applicable. If additional assumptions (continuous differenti
ability) are introduced, then we have Theorems 5-2 and 5-5 (or 5-11), which give
only sufficient conditions but are easier to apply. If we have the analyticity
assumption (the strongest possible assumption) on time-varying dynamical
equations, then we have Theorems 5-3 and 5-6 (or 5-12). Theorem 5-7 (or
5-13), which follows directly from Theorems 5-1, 5-3, and 5-4, gives the necessary
and sufficient conditions for a linear time-invariant dynamical equation to be
controllable.
The relationship between the transfer-function matrix and the linear time
invariant dynamical equation was established in this chapter. This was achieved
by decomposing a dynamical equation into four parts: (1) controllable and
observable, (2) controllable but unobservable, (3) uncontrollable and unobser
vable, and (4) uncontrollable but observable. The transfer-function matrix
depends only on the controllable and observable part of the dynamical equation.
If a linear time-invariant dynamical equation is not controllable and not
observable, it can be reduced to a controllable and observable one.
The concepts of controllability and observability are essential in the study of
Chapters 6 to 8. They will be used in the realization of a rational matrix
(Chapter 6) and the stability study of linear systems (Chapter 8). Some practical
implications of these concepts will be given in Chapter 7.
The computational problems of the various controllability and observability
conditions are also discussed. Although the conditions can be stated nicely in
terms of the ranks of [B AB ... An- 1 B] and [sl - A
B] in the con
trollability case, they are not suitable for computer computations. An efficient
and numerically stable method is introduced to transform a dynamical equation
into the form in (5-54) or (5-60), and its controllability or observability can then
be determined. The algorithm can also be used to reduce a reducible dynamical
equation to an irreducible one.
'
Before concluding this chapter, we remark on the controllability of the
n-dimensional linear time-invariant des crete-time equation
x(k
+ 1) = Ax(k) +Bu(k)
Problems
5-1
Which of the following sel
a. {t, t 2 , l, e 2t,
te t , tZel,.~ ,
{l,
b.
c. {sin t, cos t, sin 2t}

5-2
a.
Check the controllability
G;]
y
:t:J +[~]
[~
[0 1]
[Xl]
Xz
[0 Jx+[
~ x=
0
-2
-4
y=[~
c.
x=
y=[
0
0
;]x
20
-25
I!} +r
-20
1 3 0]
5-3 Show that a linear dynamic

finite II > to such that for any x o,
Hint: Use the nonsingularity of 1
5-4 Show that if a linear dynaI
at any t < to. Is it true that if a
controllable at any t> to? Why
16
In the literature, if Xo = 0, it is cal

encompasses both and does not I
ability and the condition of cont
identical.
~lCAL
PROBLEMS
EQUATIONS
Similar to Definition 5-1, we may define {A, B} to be controllable if and only if,
given anyxo and anyxb there exists an input sequence {u(k)} of finite length to
transfer Xo to Xl. 16 The condition of controllability is that rank [B AB ...
An-l BJ = n (see Problem 2-20). This condition is identical to statement 3 of
Theorem 5-7. Hence, most of the results in the time-invariant continuous-time
case are applicable to the discrete-time case without any modification. For
the time-varying case, the situation is different but is simpler. This will not be
discussed.
trollability and observa

to be controllable and
'elations among some of
dual to each other:
IS
5-8
5-16 Theorem
11
227
5-10
5-14 5-17
Problems
, derived from the con

. duality), and vice versa.
II results of this chapter.
ity), and hence they are
(continuous differenti
.5-5 (or 5-11), which give
we have the analyticity
time-varying dynamical
5-12). Theorem 5-7 (or
d 5-4, gives the necessary
ynamical equation to be
Which of the following sets are linearly independent over (
5-1
00,
oo)?
2t
a. {t, t , e , e , te }
b. {e t , te t , t 2 et , te 2t , te 3t
}
c. {sin t, cos t, sin 2t}
5-2 Check the controllability of the following dynamical equations:
. [::J ~ [~ :t:J +[~J

[::J
y
trix and the linear time

lpter. This was achieved
rts: (1) controllable and
mtrollable and unobser
transfer-function matrix
~the dynamical equation.
)t controllable and not
servabIe one.
e essential in the study of
on of a rational matrix
apter 8). Some practical
r 7.
lability and observability
1S can be stated nicely in
1- A
BJ in the con
nputations. An efficient
rm a dynamical equation
or observability can then
Lee a reducible dynamical
b.
[0 1]
=[ ~ ~ ~] + [~ ~]
x
-2
y=[~
c.
x=[~
Y
[-1
-4
-3
-;Jx
2
4
20
-25
u.
1~}+ [-~}
-20
3 0]
-1
5-3 Show that a linear dynamical equation is controllable at to if and only if there exists a
finite t 1 > to such that for any xo, there exists a u that transfers Xo to the zero state at time t 1.
Hint: Use the nonsingularity of the state transition matrix.
Show that if a linear dynamical equation is controllable at to, then it is controllable
at any t < to. Is it true that if a linear dynamical equation is controllable at to, then it is
controllable at any t> to? Why?
54
he controllability of the
lation
16
In the literature, if Xo =0, it is called reachable; if Xl =0, it is called controllable. Our definition
encompasses both and does not make this distinction. If A is singular, the condition of reach
ability and the condition of controllability are slightly different. If A is nonsingular, they are
identical.
}'
228
5-5
Is it true that p[B : AB: ... : An-IBJ =p[AB : A2B : ... : AnBJ? If not, under
what condition will it be true?
Is it possible to choose an initia

is of the form y{t) = te- t for D
Show that if a linear time-invariant dynamical equation is controllable, then it is
uniformly controllable.
5-16 Consider the dynamica

state of the equation is not kno
5-6
yiste-tfort~1'!
Check the observability of the dynamical equations given in Problem 5-2.
5-7
State (without proof) the necessary and sufficient condition for a linear dynamical
equation E to be differentially controllable and differentially observable at to.
5-8
Check the controllability of the following state equations:
5-9
a.
x = [~
b.
x=[~ ~Jx +[e~tJu
c.
x=[~ ~JX+[e~2]U
5-10
~] x +G]u
5-17 Show that the state of aI

equation can be determined i
n - 1 order. [Hint: Compute
5-18 Show that controllabilit
tions are invariant under any ei
for all t and continuously diffe]
5-19
x=[
to a controllable one.
Check the controllability and observability of.
X~[~
y
[1
-d~l +[~ ~l
5-20
Reduce the equation in 1
5-21
Reduce the folloWing dy
0 1Jx
x=
by using statement 3 of Theorems 5-7 and 5-13.

5-11 What are the controllability indices and the observability index of the equation in
Problem 5-1O?
Compute the controllability indices and the controllability index of the state equa
tion in (3-48).
5-12
5-13
to a controllable and observabl

5-22
Consider the n-dimensiol
Given a controllable linear time-invariant single-input dynamical equation
Ax
+bu
where A is an n x n matrix and b is an n x 1 column vector. What is the equivalent dy

namical equation if the basis {b, Ab, ... , AI1-1 b} is chosen for the state space? Or, equi
valently, if x Px and if p-l [b Ab ... AI1-1 bJ, what is the new state equation?
5-14 Find the dynamical equations for the systems shown in Figures 5-3, 5-4, and 5-11
and check the controllability and observability of these equations.
5-15
Reduce the dynamical ei
The rank of its controllability n
is assumed to be ni n). Let

independent columns ofU. Let
n1 x n 1 unit matrix. Show tha'
Fi
Consider the dynamical equation
[-1o 1 0] [0]
-1
0
0 x
-2
+ 1
is controllable and is zero-state

Hint: Use Theorem 5-16.
In Problem 5-22, the red

Find a method to solve P 1 in P
5-23
y=[1
1J x
AMICAL EQUATIONS
PROBLEMS
229
: ... : AnB]? If not, under
Is it possible to-choose an initial state at t = 0 such that the output of the dynamical equation
is of the form y(t) = te- t for t > O?
on is controllable, then it is
5-16
Consider the dynamical equation in Problem 5-15. It is assumed that the initial
state ofthe equation is not knowb. Is it possible to find an input u[O,oo) such that the output
y is te- t for t:::::1?
'en in Problem 5-2.

tition for a linear dynamical
observable at to.
5-17 Show that the state of an observable, n-dimensional linear time-invariant dynamical
equation can be determined instantaneously from the output and its derivatives up to
n-1 order. [Hint: Compute y(t),Ji(t), ... ,yln- 1l(t)']
is:
5-18
Show that controllability and observability of linear time-varying dynamical equa

tions are invariant under any equivalence transformation = P(t}x, where Pis nonsingular
for all t and continuously differentiable in t.
5-19
Reduce the dynamical equation

y=[llJx
to a controllable one.
5-20
Reduce the equation in Problem 5-19 to an observable one.
5-21
Reduce the follOWing dynamical equation

0
Al
in Figures 5-3, 5-4, and 5-11

lions.
Al
'0
0
[0
A2
0
1J x
to a controllable and observable equation.

5-22
Consider the n-dimensional linear time-invariant dynamical equation
dynamical equation
What is the equivalent dy

)r the state space? Or, equi
is the new state equation?
0
0
0
ility index of the equation in
)ility index of the state equa
0
0
0
FE:
Ax +Bu
y=Cx+Eu
The rank of its controllability matrix,

U = [B : AB : ... : An- 1 B]
is assumed to be nl n). Let QI be an n x nl matrix whose columns are any n 1 linearly
independentcolumnsofU. Let PI beannl x nmatrixsuchthat P 1 Ql I n1 ,whereI n1 isthe
n l x nl unit matrix. Show that the following n1-dimensional dynamical equation
FE:
Xl =P 1 AQIXl +P 1 Bu
y =CQ1X I +Eu
is controllable and is zero-state equivalent to FE. In other words, FE is reducible to
Hint: Use Theorem 5-16.
5-23 In Problem 5-22, the reduction procedure reduces to solving for PI in P1QI
Find a method to solve PI in PI-QI
Inc
230
Develop a similar statement as Problem 5-22 for an unobservable linear time-
invariant dynamical equation.
5-24
5-25
5-32
Consider two
Is the following Jordan-form dynamical equation controllable and observable?
2
0
2
2
1
o x+
x=lo
0
0
0
0
1
0
0
2
1
2
1
Y=l:
x=[j
Y=
["
2 1IU
5-33 A q x p rational matrix

matrix GLI(s) such that Gu(s)
pG(s) = p in [R(s).
5-34
~}
C12
C13
C14
C21
C22
C23
C24
Cl:J
C31
C32
C33
C34
C35
C25
b 21
b 31
b41
b 51
Let P be a nonsingular 1
PB=[l
0] [" b'~
~1)
0 0
o .
o X+
o
(s
are they output controllable?
Is it possible to find a set of bij and a set of Cu such that the following Jordan-form
equation
0
G1 =
5-26
1
0
0
0
s+1
0 0
0 1
0
0
systems'~
b 22
b32
b42
b5
where Bl is an nl x P matrix al
sions (n n l ) x nl and (n - n l )
if and only if {A22' A2d is con1
5-35
pU
Show that
p[B AB ... An-IB]:
where linearly independent col

Hint: Premultiply T# by
is controllable? Observable?
Show that A is cyclic (see Problem 2-45) if and only ifthere exists a vector b such that
{A, b} is controllable. (Hint: Use Corollary 5-21.)
5-27
5-28 Show that if (A, B} is controllable and A is cyclic, then there exists a
vector r such that {A, Br} is controllable. (Hint: Use Theorem 5-21.)
5-36 Show that {A, C} is obsel

are, respectively, n x nand q x 11
pose of C.
x 1 column
Show that a necessary condition for {A, B} where B is an n x p matrix, to be control

lable is p 2?:.m, where m is the largest number of Jordan blocks associated with the same
eigenvalue of A. (See Problem 5-26 and use Theorem 5-21.)
5-29
5-30 In Corollary 5-22 we have ,that a dynamical equation with E = 0 is output

controllable if and only if p[CB: CAB: ... : CAn-1B] =q. Show that if E =1=0, then the
dynamical equation is output controllable if and only if
p[CB: CAB:"': CAn- 1 B: E] =q

5-31 Consider a linear time-invariant dynamical equation with .E =0. Under what
condition on C will the (state) controllability imply output contr~bi1ity?
Show that lA, B} is con'

that is orthogonal to all colum
left eigenvector a: of A such that
5-37
A
This is called the Popov-Be/evit,
of Theorem 5-7.)
PROBLEMS
MICAL EQUATIONS
unobservable linear time-
5-32
231
Consider two systems with the transfer-function matrices
(; _[S~l
1
1-
(s
+ 1)
(S~~2;s3~1)l
(;z
(s+2)
(s
[(S~2)l
(s+1)
+ 3)(s + 1)
(s
+ 2)
are they output controllable? Output function controllable?

5-33 A q x p rational matrix G(s) is said to have a left inverse if there exists a rational
matrix Gu(s) such that Gu(s)G(s) = Ip- Show that G(s) has a left inverse if and only if
p(;(s) = pin Il!(s).
5-34
Let P be a nonsingular matrix such that

PB=[B01 ]
t the following Jordan-form
PAP- 1
=rLAll
AZl
A12]
A22
where B1 is an n1 x P matrix and pB = pBl nl' The matrices A21 and A22 have dimen
sions (n - nd x nl and (n - nl) x (n - nd, respectively. Show that {A, B} is controllable
if and only if {An' Az d is controllable.
5-35
Show that
pU =p[B
AB ... An- 1B] total number of linearly independent B columns ofTI'

in (5-94)
where linearly independent columns of T I' are to be searched in order from left to right.
Hint: Premultiply TJ.! by
,re exists a vector b such that
5-36 Show that (A, C} is observable if and only if {A, C*C} is observable, where A and C
are, respectively, n x nand q x n constant matrices and C* is the complex conjugate trans
pose of C.
there exists a p x 1 col umn
m 5-21.)
n n x p matrix, to be control
ks associated with the same
ltion with E = is output

Show that if E '" 0, then the
tl
with E =0.
ontf(~bility?
Under what
5-37 Show that {A, B} is controllable if and only if there exists no left eigenvector of A
that is orthogonal to all columns of 8, that is, there exist no eigenvalue I. and nonzero
left eigenvector IX of A such that
and
IXB=O
This is called the Popov-Belevitch-Hautus test in Reference S125. (Hint: See statement 4
of Theorem 5-7.)

Chen ch5

Enviado por

Dados do documento

Título original

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

Chen ch5

Enviado por

Direitos autorais:

Formatos disponíveis

~

CONTROLLABILITY AND OBSERVABILITY OF LINEAR DYNAMICAL EQUATIONS

The dynamical equation E is said to be uniformly controllable if and only if there

0< a3(uJI ;s;cJ)(t +U,' t)W(t, t +u,)cJ)*(t +U,' t) ;s;a 4 (u,)I

Figure 5-5 The function b(t).

By A> B, we mean that the matrix (A - B) is a positive definite matrix (see

achieved in the time interval U c :

Consider the one-dimensional linear dynamical equation

where x is the n x 1 state vector,

e- 2t dt =0.5e- 2t(l- e- 211,) > a I (uc;)

A remark is in order concerning controllability, differential controllability,

The n-dimensional linear time-inva

1. All rows of e-AtB (and conseqt:

Although all the entries of A and Bare real I

CONTROLLABILITY OF LINEAR DYNAMICAL EQUATIONS

y controllable if and only if there

.tion matrix and W is as defined

is a positive definite matrix (see

Figure 5-5 The function b(t).

lical equation is instantaneously

al equation is not instantaneous

lility, differential controllability,

where x is the n x 1 state vector, u is the p x 1 input vector: A and Bare n x n

CONTROLLABILITY OF LINEAR DYNAMICAL EQUATIONS

v controllable if and onl y if there

(J'c such that

:;; (l2((J' c)1

lion matrix and W is as defined

ls a positive definite matrix {see

:ontrollability both imply con

Figure 5-5 The function b(t}.

cal equation is instantaneously

I equation is not instantaneous

Iity, differential controllability,

where x is the n x 1 state vector, u is the p x 1 input vector; A and Bare n x n

1. All rows of e-AtB (and consequently of eAtB) are linearly independent on

CONTROLLABILITY AND OBSERVABILITY OF LINEAR DYNAMICAL EQUATIONS

2. The controllability grammian

W,,~ J~ e"BB'e A ', dT

Consider the inverted pendulL

is nonsingular for any t > 0. 7

AB: A2B : ... : A"-IB]

p[ e- AtB : - e- AtAB : . . . : ( 1)"- 1 e - AtA IB : ...]

From the Cayley~-Hamilton theorem, we know that Am with m ~n can be written

Since the Laplace transform is a one-to-one linear operator, if the rows of

The matrix is in fact positive definite. See Problem E-11.

This condition implies that lsI

A) and B are left coprime. See Appendix G.

This matrix can be readily shO\

Consider the platform system st

This is the state-variable descrip

hence the state equation is contr

CONTROLLABILITY OF LINEAR DYNAMICAL EQUATIONS

)r every}, in C), the n x (n

from Theorems 5-1 and

,Am with m ;? n can be written

!ar operator, if the rows of

(page 206). Q.E.D.

This is the state-variable description of the system.

hence the state equation is controllable. Consequently, the displacements can

CONTROLLABILITY AND OBSERVABILITY OF LINEAR DYNAMICAL EQUATIONS

be brought to zero in 2 seconds by applying a proper input u. Using Equations

for t in [0, 2]. If a force of the