EE650 - Basics of Modern Control Systems

Solution for Assignment - 6

1. To evaluate the state-transition matrix Φ(t, τ ) for the given system with:
 
cos(t) 0
A(t) = ,
0 1

the state-transition matrix satisfies the differential equation:

∂
Φ(t, τ ) = A(t)Φ(t, τ ), Φ(τ, τ ) = I.
∂t
The state-transition matrix is expressed as:
Rt
A(s) ds
Φ(t, τ ) = e τ .

Thus: Z t  
sin(t) − sin(τ ) 0
A(s) ds = .
τ 0 t−τ

Exponentiate the resulting matrix to compute Φ(t, τ ):

 sin(t)−sin(τ ) 
Rt
e 0
Φ(t, τ ) = e τ A(s) ds = .
0 et−τ

The state-transition matrix is:  sin(t)−sin(τ ) 

e 0
Φ(t, τ ) = .
0 et−τ

2. similar to question 3.
3. The given system is:  
0 0
ẋ(t) = x(t)
t 0

or equivalently:
ẋ1 (t) = 0,
ẋ2 (t) = tx1 (t).

The solution to ẋ1 (t) = 0 with initial condition x1 (0) is:

x1 (t) = x1 (0).

Substitute x1 (t) = x1 (0) into ẋ2 (t) = tx1 (t):

ẋ2 (t) = tx1 (0).

Integrating both sides: Z t

x2 (t) = τ x1 (0) dτ + x2 (0).
0

Evaluate the integral:

1 2
x2 (t) = t x1 (0) + x2 (0).
2

1
 Thus, the solution for x(t) is:
   
x1 (t) x1 (0)
x(t) = = 1 2 .
x2 (t) 2 t x1 (0) + x2 (0)

 
1
- If x(0) = , then:
0
 
1
x(t) = 1 2 .
2t
 
1
- If x(0) = , then:
2

1
x(t) = 1 2 .
2t + 2

The two initial states are linearly independent. Therefore, the fundamental matrix Φ(t) is given by:
 
1 1
Φ(t) = 1 2 1 2 .
2t 2t + 2

4. similar to question no 6
5. (a) The matrix A has eigenvalues/eigenvectors given by
      
0 1 v1 v1 1
λ1 = 2, (A − 2I)v = 0 ⇒ =0 ⇒ =α , ∀α ∈ R,
0 −4 v2 v2 0
      
4 0 v1 v1 1
λ2 = −2, (A + 2I)v = 0 ⇒ =0 ⇒ =α , ∀α ∈ R.
0 1 v2 v2 −4

Thus, A can be diagonalized as

     
−1 2 0 1 1 −1 1 1/4
A = T DT , D= , T = , T = .
0 −2 0 −4 0 −1/4

Therefore,    2t  
tA tD −1 1 1 e 01 1/4
e = Te T =
0 −4 0 e−2t 0 −1/4

e−2t e2t −e−2t

 2t    
e 1 1/4 e2t
= = 4 .
0 −4e−2t 0 −1/4 0 4e −2t

Alternatively, one could compute etA using

" #
s−2 −1
tA −1 −1 −1 (s−2)(s+2) (s−2)(s+2)
e =L {(sI − A) }=L s+2
0 (s−2)(s+2)

e2t −e−2t
 
e2t
=⇒ 4
−2t .
0 e

(b) The system is unstable because A has one eigenvalue with a positive real part.
6. (a) Transfer Matrix Computation
The transfer matrix is given by:
  −1  
  −2 0 0 1
Ĝ(s) = 1 1 1 sI −  0 0 1 1 .
0 0 0 0
 Simplifying:
 −1  
  s+2 0 0 1
Ĝ(s) = 1 1 1  0 s −1 1 .
0 0 s 0

Let:    −1  
z1 s+2 0 0 1
z2  =  0 s −1 1 .
z3 0 0 s 0

We solve:     
s+2 0 0 z1 1
 0 s −1 z2  = 1 .
0 0 s z3 0

Equivalently:
z3 = 0,
1
z2 = ,
s
1
z1 = .
s+2
Thus:  
  z1 1 1 2s + 2
Ĝ(s) = 1 1 1 z2  = + +0= .
s+2 s s(s + 2)
z3

(b) The matrix has a 2 × 2 Jordan block with zero eigenvalue, which makes the system unstable.
(c)The system is not BIBO stable because it has a pole with zero real part.

7. First we will put the system into state variable form.

   
0 1 0
ẋ = x+ u
2 −1 1
 
y = −1 1 x

Finding the eigenvalues of this system, we determine that σ(A) = {1, −2}. From these eigenvalues, we
see that this system is neither asymptotically nor Lyapunov stable. We will now check to see whether it
is BIBO stable. It is an easy exercise to check that

+ 13 e−2t − 13 e−2t
2 t 1 t

e At
= 3e 3e
2 t
3e − 23 e−2t 1 t
3e + 23 e−2t

and that CeAt B = e−2t

Therefore,
Rt Rt
−∞
||CeA(t−τ ) B||dτ = −∞ ||e−2(t−τ ) ||dτ = 1
2
which is bound =⇒ The system is therefore BIBO stable.