ELEC-C8201: Control Theory and Automation

Exercise 8 - Solutions

The problems marked with an asterisk (?) are not discussed during the exercise session. The
solutions are given in MyCourses and these problems belong to the course material.

1. Consider the single-input, single-output system described by

ẋ(t) = Ax(t) + Bu(t),

y(t) = Cx(t)

Determine whether the system is controllable and observable when:

   
0 1 0  
a) A = , B= , C= 0 2 .
0 −3 1

   
−10 0 0  
b) A = , B= , C= 1 0 .
0 −2 2

   
0 1 −1  
c) A = , B= , C= 1 0 .
−1 −2 2

Solution.

a) Controllability:
 
 0 1
 0 1
Pc = B AB = ⇒ det(Pc ) = = −1 6= 0 (→ controllable)
1 −3 1 −3

Observability:

  
C 0 2
Po = = ⇒ det(Po ) = 0 (→ not observable)
CA 0 −6

b) Controllability:
 
  0 0
Pc = B AB = ⇒ det(Pc ) = 0 (→ not controllable)
2 −4

Observability:
   
C 1 0
Po = = ⇒ det(Po ) = 0 (→ not observable)
CA −10 0

Note that in this case, the input is only on x2 and therefore x1 cannot be controlled.
Given, however, that A is diagonal, we can easily see that x1 goes to zero for whatever
input and hence the system is stabilizable. Similarly, If state x2 goes to zero, the
system is detectable.
 c) Controllability:
 
  −1 2 −1 2
Pc = B AB = ⇒ det(Pc ) = = 3 − 4 = −1 6= 0 (→ controllable)
2 −3 2 −3

Observability:
  
C 1 0
Po = = ⇒ det(Po ) = 1 (→ observable)
CA 0 1

2. Consider the second-order system

   
1 −1 k
ẋ(t) = x(t) + 1 u(t),
−1 1 k2
 
y(t) = 1 0 x(t)

For what values of k1 and k2 is the system completely controllable?

Solution.
 
k k1 − k2
Pc = B AB = 1
 
k2 −k1 + k2

Therefore, the determinant is given by

k1 k1 − k2
det(Pc ) = = −k1 (k1 − k2 ) − k2 (k1 − k2 ) = −(k1 − k2 )(k1 + k2 )
k2 −k1 + k2

6 |k2 | (or k12 6= k22 ).

So, the system is completely controllable if and only if |k1 | =

3. Consider the third-order system

   
0 1 0 0
ẋ(t) = 0
 0 1 x(t) + 0 u(t),
 
−8 −5 −3 4
 
y(t) = 2 −4 0 x(t)

a) Verify that the system is observable.

b) Determine the observer gain matrix required to place the observer poles at s1,2 = −1 ± j
and s3 = −5.
 Solution.
a) Observability:
   
C 2 −4 0
Po =  CA  =  0 2 −4 ⇒ det(Po ) = 728 6= 0 (→ observable)
CA 2 32 20 14
b) The desired characteristic equation is given by
p(s) = [s − (−1 + j)][s − (−1 − j)](s + 5)
= [(s + 1)2 + 1](s + 5)
= (s2 + 2s + 2)(s + 5)
= s3 + 7s2 + 12s + 10
1st way: Using the Ackermann’s formula for the observer gain matrix,
 
0.14
T
L = p(A)Po−1 0 0 . . . 1 = . . . = −0.93
 

0.79
where p(A) is given by
p(A) = A3 + 7A2 + 12A + 10I.
2nd way: Using the determinant:
     
s 0 0 0 1 0 l1  
|sI − (A − LC)| = 0 s 0 − 0
   0 1 + l2  2 −4 0
 
0 0 s −8 −5 −3 l3

4. Suppose that the vector differential equation describing the inverted pendulum is given by
   
0 1 0 0 0
0 0 −1 0 1
ẋ(t) = 
0 0 0
 x(t) +   u(t)
1 0
0 0 9.8 0 −1
Assume that all state variables are available for measurement and use state variable feedback.
Place the system characteristic roots at s − −2 ± j ,−5, and −5.

Solution. Consider the state variable feedback law u = Kx. Using the Ackermann’s formula
 
K = −14.2045 −17.0455 −94.0045 −31.0455

5. A system is represented by the differential equation

d2 y dy du
+2 +y = +u
dt2 dt dt
where y = output and u = intput.
 a) Develop a state variable representation and show that it is a controllable system.

b) Define the state variables as x1 = y and x2 = dy/dt − u, and determine whether the
system is controllable.

Solution.
a) In the first case:    
0 1 0
A= , B=
−1 −2 1
Therefore,
 
  0 1 0 1
Pc = B AB = ⇒ det(Pc ) = = −1 6= 0 (→ controllable)
1 −2 1 −2
b) In the second case:    
0 1 1
A= , B=
−1 −2 −1
Therefore,
 
  1 −1 1 −1
Pc = B AB = ⇒ det(Pc ) = =1−1=0 (→ not controllable)
−1 2 −1 1
Note: the controllability of a system depends on the definition of the state variables!

6. Consider the second-order system

   
1 2 −5
ẋ(t) = x(t) + u(t),
−6 12 1
 
y(t) = 4 −3 x(t)
a) Verify that the system is observable and controllable.

b) Design a full-state feedback law and an observer by placing the closed-loop system poles
at s1,2 = −1 ± j and the observer poles at s1,2 = 12.

Solution.
a) Controllability:
 
  −5 −3
Pc = B AB = ⇒ det(Pc ) = −87 6= 0 (→ controllable)
1 18
Observability:

  
C 4 −3
Po = = ⇒ det(Po ) = 242 6= 0 (→ observable)
CA 22 44
 
b) The controller gain matrix K = 3.02 6.11 places the closed-loop system poles at
 T
s1,2 = −1 ± j, and the observer gain matrix L = 2.38 −1.16 observer poles at
s1,2 = 12.