Instructions for filling up the exam sheet

This solutions sheet MAY be compiled with a pencil

For the exam, ONLY this solutions sheet should be turned in. The exam will be graded using this sheet only.
In case of lack of time, students may turn in their exam without (or only partially) compiling the solutions sheet.
This will however result in a −3 points penalty in the final grade.
Exam of Automatic Control (9 credits) - February 31, 2031 1
Surname: Immatriculation number:
Name: Place and date of birth:

Exercise ZSE (5 points) Consider the two continuous-time (or discrete-time) LTI systems characterized by the following
matrices:
   
1 1 −1 0  
A1 = 0, B1 = 1, C1 = −4, D1 = 1; A2 = , B2 = , C2 = 4 4 , D2 = 1
2 −1 1 −1

and say whether they are zero state equivalent and whether each of them is reachable, observable, stabilizable and
detectable.

Solution to exercise ZSE

System 1 is System 2 is
G1 (s)|G1 (z) =
 Reachable  Reachable The two systems:
 Stabilizable  Stabilizable  Are ZSE
 Observable  Observable  Are NOT ZSE
G2 (s)|G1 (z) =
 Detectable  Detectable
 
3 −1
Exercise EXP (6 points) Given matrix A = , and knowing that with the similarity transformation T =
  2 0
−1 1
one obtains the Jordan for J = T AT −1 of A, compute the matrix exponential eAt or the matrix power At
−2 1
using the Jordan form of A.
Solution to exercise EXP

J= eJt |J t = eAt |At =


ẋ = Ax + Bu
Exercise SS (4 points) Write the matrices (A, B, C, D) of a realization of the transfer function
y = Cx + Du

s2 − 3
G(s) = .
(s + 1)2

Moreover, indicate whether the realization is stable, asymptotically stable or unstable.

Solution to exercise SS  
    The state-space realization is
A B    Asymptotically Stable
= ;
C D 


  Stable (not asymptotically)
 Unstable


ẋ = Ax + Bu
Exercise GS (5 points) Given the continuous-time LTI system characterized by the following matri-
y = Cx + Du
ces:    
1 4 −2    
A= , B= , C = −3 1 , D = 1 ,
−1 −3 1
compute the corresponding transfer function G(s) fromu to y. 
−1 1
Moreover, knowing that the Jordan form of A is J = , compute a similarity transformation matrix T such
0 −1
that, using the transformed coordinates x̄ = T x, we have Ā = T AT −1 = J.
Finally, compute matrices (Ā, B̄, C̄, D̄) of the system in the transformed coordinates x̄.
Exam of Automatic Control (9 credits) - February 31, 2031 2

Solution to exercise GS  
 
   
Ā B̄  
G(s) = T = ; = .
 
C̄ D̄ 




Exercise LYAP (4 points) Consider the following hybrid system with x = [ xx12 ] and the attractor A = {0}:

x2 − x31
 
2 2
 x ∈ C = {x : x1 ≥ x2 }, ẋ = f (x) = − 1 x1



2
 
 2 2 + x 1
 x ∈ D = {x : x1 ≤ x2 }, x = g(x) = 1


2 x1

Using the Lyapunov function candidate (this function trivially satisfies (S’)):
1 2
V (x) = x + x22 ,
2 1
verify whether conditions (F”), (J”) hold (which implies UGS). Also verify if (F’), (J’) hold (which implies UGAS) or
not. If (F”), (J”) hold but (F’), (J’) do not hold, say whether (I) holds (which would imply UGAS) or not.

Solution to exercise LYAP

V̇ (x) = ∆V (x) =

(F”) and (J”) are satisfied:  YES,  NO.

(F’) and (J’) are satisfied:  YES,  NO.

(I) is satisfied:  YES,  NO. (only if applicable)

Exercise LIN (6 points) Consider the continuous-time or discrete-time plant:

ẋ1 = −x1 + 3(x2 − 1),

ẋ2 = x22 − 1

compute all the equilibrium points and the linearization of the dynamics about each one of them.
Based on the linearization, say whether the specific equilibrium is: 1) exponentially stable; 2) unstable; 3) nothing can
be concluded.
Solution to exercise LIN

∂f
= Equilibria: xa = xb = ...
∂x

Equilibrium  xa   xb   xc   xd 
       
       
Linearization matrix A 















       

Exponentially Stable    
Unstable    
Can’t say    
Exam of Automatic Control (9 credits) - February 31, 2031 3

 
−7 −5
Exercise MEC (5 points) Consider the continuous-time linear system characterized by A = , eAt =
8 6
 −2t
− 5et 5e−2t − 5et
  
1 8e 1
and B = and compute, whenever possible, the control law taking values in the
3 8e − 8e−2t 8et − 5e−2t
t
−1
set [t0 , t1 ] = [0, 2] and taking the state from x(t0 ) = x(0) = 0, to x(t1 ) = x(2) = xf inal , where xf inal takes the following
values      
−1 1 1
a) xf inal = xa = , b) xf inal = xb = , c) xf inal = xc = .
1 1 0
In particular, for cases a), b) and c) indicate whether such a control exists and, if it exists, give its explicit form.

Solution to exercise MEC

>
eAt BB > eA t
= WR (t0 , t1 ) =

Case a):  A solution DOES NOT EXIST;  A solution EXISTS and is

η1 = ua (t) =

Case b):  A solution DOES NOT EXIST;  A solution EXISTS and is

η1 = ub (t) =

Case c):  A solution DOES NOT EXIST;  A solution EXISTS and is

η1 = uc (t) =

Exercise KAL (5 points) Consider the continuous-time or discrete-time linear dynamics described by:
 
  1 1 −2 2
A B  0 1 0 1 
=  −1

C 2 0 1 
0 1 0

and compute its Kalman decomposition. Specify if the system is controllable, observable, stabilizable and detectable.
(Note: This exercise may be alternatively formulated for the controllable or observable decompositions only.)

Solution to exercise KAL (Only fill up the relevant parts)

R= 1) Basis of Im R ∩ Ker O = 2) Completion of Im R =

O= 3) Completion of Ker O = 4) Completion of T =

 
 
   
Ā B̄
 
T = T −1 = =
 
C̄

 
 
 

The system is :  Reachable,  Stabilizable,  Observable,  Detectable

Exam of Automatic Control (9 credits) - February 31, 2031 4

  
1 0  
Exercise EIG (4 points) Given a system characterized by the matrices (A, C) = , 0 1 , determine a gain
3 3
matrix L assigning the eigenvalues of the closed-loop matrix A − LC exactly in λ1 = −1, λ2 = −1.

Solution to exercise EIG

pA (s) = pAcl (s) =

R ( or R̄) = M ( or M̄ ) = K( or L) =

R−1 ( or R̄−1 ) = M −1 ( or M̄ −1 ) =