9-95

Design Methods

Solving the preceding equations for each i and substituting them into Equation 9.136, we obtain
1 = [n n1 1 ] T 1
K = KT
= [n an |n1 an1 | |2 a2 |1 a1 ] T

(9.138)

Thus, if the system is completely state controllable, all eigenvalues can be arbitrarily placed by choosing
matrix K according to Equation 9.138 (sufficient condition).
We have thus proved that the necessary and sufficient condition for arbitrary pole placement is that
the system be completely state controllable.

9.6.4 Design Steps for Pole Placement

Suppose that the system is defined by
x = Ax + Bu
and the control signal is given by
u = Kx
The feedback gain matrix K that forces the eigenvalues of A BK to be 1 , 2 , . . . , n (desired values)
can be determined by the following steps. (If i is a complex eigenvalue, then its conjugate must also be
an eigenvalue of A BK.)
Step 1
Check the controllability condition for the system. If the system is completely state controllable, then use
the following steps.
Step 2
From the characteristic polynomial for matrix A:
|sI A| = sn + a1 sn1 + + an1 s + an
determine the values of a1 , a2 , . . . , an .
Step 3
Determine the transformation matrix T that transforms the system state equation into the controllable
canonical form. (If the given system equation is already in the controllable canonical form, then T = I.)
It is not necessary to write the state equation in the controllable canonical form. All we need here is to
find the matrix T. The transformation matrix T is given by Equation 9.129, or
T = MW
where M is given by Equation 9.130 and W is given by Equation 9.131.
Step 4
Using the desired eigenvalues (desired closed-loop poles), write the desired characteristic polynomial
(s 1 )(s 2 ) (s n ) = sn + 1 sn1 + + n1 s + n
and determine the values of 1 , 2 , . . . , n .
Step 5
The required state feedback gain matrix K can be determined from Equation 9.138, rewritten thus:
K = [n an |n1 an1 | |2 a2 |1 a1 ]T 1

9-96

Control System Fundamentals

9.6.5 Comments
Note that if the system is of lower order (n 3), then direct substitution of matrix K into the desired
characteristic polynomial may be simpler. For example, if n = 3, then write the state feedback gain
matrix K as
K = [k1 k2 k3 ]
Substitute this K matrix into the desired characteristic polynomial |sI A + BK| and equate it to
(s 1 )(s 2 )(s 3 ), or
|sI A + BK| = (s 1 )(s 2 )(s 3 )
Since both sides of this characteristic equation are polynomials in s, by equating the coefficients of the
like powers of s on both sides it is possible to determine the values of k1 , k2 , and k3 . This approach is
convenient if n = 2 or 3. (For n = 4, 5, 6, . . . , this approach may become very tedious.)
There are other approaches for the determination of the state feedback gain matrix K. In what follows,
we present a well-known formula, known as Ackermanns formula, for the determination of the state
feedback gain matrix K.

9.6.6 Ackermanns Formula

Consider the system given by Equation 9.126, rewritten thus:
x = Ax + Bu
We assume that the system is completely state controllable. We also assume that the desired closed-loop
poles are at s = 1 , s = 2 , . . . , s = n .
Use of the state feedback control
u = Kx
modifies the system equation to
x = (A BK)x

(9.139)

Let us define
A = A BK
The desired characteristic equation is

|sI A + BK| = |sI A|

= (s 1 )(s 2 ) (s n )
= sn + 1 sn1 + + n1 s + n = 0
Since the CayleyHamilton theorem states that A satisfies its own characteristic equation, we have
= A n + 1 A n1 + + n1 A + n I = 0
(A)

(9.140)

We utilize Equation 9.140 to derive Ackermanns formula. To simplify the derivation, we consider the
case where n = 3. (For any other positive integer n, the following derivation can be easily extended.)

9-97

Design Methods

Consider the following identities:

I =I
A = A BK
A 2 = (A BK)2 = A2 ABK BK A
A 3 = (A BK)3 = A3 A2 BK ABK A BK A 2
Multiplying the preceding equations in order by 3 , 2 , 1 , 0 (where 0 = 1), respectively, and adding
the results, we obtain
3 I + 2 A + 1 A 2 + A 3
+ A3 A2 BK ABK A BK A 2
= 3 I + 2 (A BK) + 1 (A2 ABK BK A)

(9.141)

= 3 I + 2 A + 1 A2 + A3 2 BK 1 ABK 1 BK A A2 BK ABK A BK A 2
Referring to Equation 9.140, we have
=0
3 I + 2 A + 1 A 2 + A 3 = (A)
Also, we have
3 I + 2 A + 1 A2 + A3 = (A) = 0
Substituting the last two equations into Equation 9.141, we have
= (A) 2 BK 1 BK A BK A 2 1 ABK ABK A A2 BK
(A)
= 0, we obtain
Since (A)
+ A2 BK
(A) = B(2 K + 1 K A + K A 2 ) + AB(1 K + K A)

2 K + 1 K A + K A 2

= [B | AB | A2 B]
1 K + K A
K

(9.142)

Since the system is completely state controllable, the inverse of the controllability matrix
[B | AB | A2 B]
exists. Premultiplying the inverse of the controllability matrix to both sides of Equation 9.142, we obtain

2 K + 1 K A + K A 2

[B | AB | A2 B]1 (A) =
1 K + K A
K
Premultiplying both sides of this last equation by [0 0 1], we obtain

2 K + 1 K A + K A 2
=K
[0 0 1][B | AB | A2 B]1 (A) = [0 0 1]
1 K + K A
K

which can be rewritten as

K = [0 0 1][B | AB | A2 B]1 (A)

This last equation gives the required state feedback gain matrix K.

9-98

Control System Fundamentals

For an arbitrary positive integer n, we have

K = [0 0 0 1][B | AB | | An1 B]1 (A)

(9.143)

Equations 9.139 to 9.143 collectively are known as Ackermanns formula for the determination of the
state feedback gain matrix K.

Example 9.2:
Consider the system defined by
x = Ax + Bu

where

0
A= 0
1

1
0
5

0
B = 0
1

0
1 ,
6

By using the state feedback control u = Kx, it is desired to have the closed-loop poles at s = 2 j4,
and s = 10. Determine the state feedback gain matrix K .
First, we need to check the controllability of the system. Since the controllability matrix M is given
by

0
0
1
1 6
M = [B | AB | A2 B] = 0
1 6 31
we find that det M = 1 and therefore rank M = 3. Thus, the system is completely state controllable
and arbitrary pole placement is possible.

Next, we solve this problem. We demonstrate each of the three methods presented in this chapter.
Method 1
The first method is to use Equation 9.128. The characteristic equation for the system is


 s 1
0 

|sI A| =  0
s 1 
 1
5 s+6 
= s3 + 6s2 + 5s + 1
= s3 + a1 s2 + a2 s + a3 = 0
Hence,
a1 = 6,

a2 = 5,

a3 = 1

The desired characteristic equation is

(s + 2 j4)(s + 2 + j4)(s + 10) = s3 + 14s2 + 60s + 200
= s3 + 1 s2 + 2 s + 3
=0
Hence,
1 = 14,

2 = 60,

3 = 200

Referring to Equation 9.138 we have

K = [3 a3 | 2 a2 | 1 a1 ]T 1