Karen Reyes ECE 602

FunWork 4 Tuesday, November 05, 2019

Problem 1
Consider the following diagram of an armature controlled DC motor, where the system
parameters are: Ra = 1, La ≈ 0H, Kb = 5V /rad/sec, Ki = 5N m/A, and the equivalent
rotational inertia, Ieq = 0.1kgm2 . The gear inertia and friction are negligible. Let x1 =
θ, x2 = θ̇, and u = ea .

Figure 1: Block diagram of a DC motor system for Problem 1.

1. Construct the state space model of the DC motor system

• Applying Kirchhoff’s voltage law to the armature circuit:
dia dθ
ea = Ra ∗ ia + La ∗ + kb ∗ (1)
dt dt
• Since La is approximate zero:
dθ
ea = Ra ∗ ia + kb ∗ (2)
dt
• Applying also Newton’s second law of motion:
d2 θ dθ
T = Ieq ∗ +f ∗ (3)
dt dt
• Since f is negligible:
d2 θ
T = Ieq ∗ (4)
dt
• Appliying Laplace transform in equations (2) and (4), and solving for I(s):
Ea − kb sθ(s)
I(s) = (5)
Ra
Ieq s2 θ(s)
I(s) = (6)
ki

1
Karen Reyes ECE 602
FunWork 4 Tuesday, November 05, 2019

θ(s)
• Putting together equation (5) and (6), and solving for Ea
:

θ̇(s) ki
= 2
(7)
Ea Ra Ieq s + ki kb s
ki
θ̇(s) R I
= 2 a keqi kb s (8)
Ea s + Ra Ieq

θ̇(s) 50
= 2 (9)
Ea s + 250s
• The state space model of the motor:
   
0 1 0
ẋ = x+ u (10)
0 −250 50
 
y = 10 x (11)

2. • Let x(0) = [10 1]T . Find a control u that brings the system to x(2) = [0 0]T :

• In order to find the control u, fist W (t0, t1) was calculated using MatLab and the
following formula: Z t1
0
W (t0 , t1 ) = eAt BB 0 eA τ dxτ (12)
t0

• The resulting matrix was :

 
0.0798 0.0200
W (t0 , t1 ) = (13)
0.0200 5.0000

• The control u was founded using Matlab and the result from equation (13).

u(t) = −B 0 eA(t1 −t) W −1 (t0 , t1 )[eAt1 x0 − x1 ] (14)

2495 ∗ e250t−500 5 e−500 2501 1250 ∗ e250t−500 625
u(t) = −e−500 ∗ ( − )−( − )( − )
249 498 250 250 249 249
(15)
• The Matlab script of the previous calculations:

2
Karen Reyes ECE 602
FunWork 4 Tuesday, November 05, 2019

3. Verify your calculations with simulations. Plot x1 , x2 , and u versus time:

• Plot x1 , x2 versus time:

Figure 2: Plot x1 , x2 versus time.

3
Karen Reyes ECE 602
FunWork 4 Tuesday, November 05, 2019

• Plot u versus time:

Figure 3: Plot u versus time

Problem 2
Consider a dynamical system modeled by
   
−1/2 1/2 −1
ẋ = x+ u (16)
−1/2 1/2 1
 
y= 1 1 x (17)

Determine x(t) over the time interval [0, 10] knowing that u(t) = 1 for t ≥ 0 and y(t) = t2 +2t
for t ≥ 0.

1. Establishing the formula to determine the solution for x(t)

Z t
A(t−t0 )
x(t) = e x(t0 ) + eA(t−τ ) Bu(τ )dτ (18)
t0

2. Calculate the parameters used in formula (14):

Z t1
At0 −1 AT t0 T (t−t )
x(t0 ) = e V (t0 , t1 )e eA 0
C T g(t)dt (19)
t0

4
Karen Reyes ECE 602
FunWork 4 Tuesday, November 05, 2019

Where: Z t1
−1 T (t)
V (t0 , t1 ) = eA C T CeAt dt (20)
t0

Using MatLab the equation was solve and the result was as follow:
 730 −970 
−1 3 3
V (t0 , t1 ) = −970 1330 (21)
3 3

and for g(t): Z t

g(t) = y(t) − C eA(t−τ ) bu(τ )dτ (22)
0

the answer for equation (18) using MatLab tools:

t2 − 2t t2 + 2t
g(t) = 2t − ( )−( ) + t2 (23)
2 2
g(t) = 2t (24)

3. Then, plogging the result from equation (17) and (20) into (15) the result is:
 
−1
x(t0 ) = (25)
1

4. Finally, using the obtained result in (21) into (14) and with the help of MatLab, the
final result was obtained:

t2 −2t
   t2 
t+ 2
−1 2
−1
x(t) = t2 +2t = t2 (26)
t+ 2
+1 2
+ 2t + 1

5. The MatLab scrip is in the follow picture, to show the process in the previous equations:

5
Karen Reyes ECE 602
FunWork 4 Tuesday, November 05, 2019

Problem 3
Transform the system into its controller form:
   
0 0 1 0 0
0 0 0 1 0
ẋ = 
2
x +  u (27)
1 0 0 1
1 1 0 0 0

1. In order to calculate the controller form, I first started calculating the controllability
matrix using the following formula and implementing it in Matlab.
 
0 1 0 −2
  0 0 0 1
b Ab A2 b A3 b =   (28)
1 0 −2 0 
0 0 1 0

2. Extract the last row of the controllability matrix which is equal to q1 and will be used
t calculate the transformation matrix in the following step:

 
q1 = 0 0 1 0 (29)

3. The transformation matrix was calculated using the results from equation (25) and the
following formula:
   
q1 0 1 0 0
 q1 A  0 0 0 1 
T =
q1 A2  = 1 −1 0 0 
   (30)
q1 A3 0 0 1 −1

4. Then, calculating the matrix A in its new coordinates:

 
0 1 0 0
0 0 1 0
à = T AT −1 =
0
 (31)
0 0 1
−1 0 −3 0

5. Calculating the matrix b in its new coordinates:

 
0
0
b̃ = T b =  
0 (32)
1

6
Karen Reyes ECE 602
FunWork 4 Tuesday, November 05, 2019

6. Putting together the matrix in controller form:

   
0 1 0 0 0
0 0 1 0 0
 
(Ã, b̃) = 
0 ,  (33)
0 0 1 0
−1 0 −3 0 1

7. The Matlab script used in order to find the solution to this problem:

Problem 4
For the system construct a state variable transformation so that in the new coordinates the
controllable part is separated and distinct from the uncontrollable part.
   
−1 1 0 1 1 1
1 1 0 0 0 0
ẋ = 
0
x +  u (34)
0 4 0 0 0
1 1 −1 2 0 1

1. The controllability matrix of the pair (A,B) is found knowing the formula ant applying
the Matlab command ctrb:

 
CM = U = B AB A2 BA3 B (35)
 
1 1 −1 0 3 4 −1 4
0 0 1 1 0 1 3 5
U =
0 0 0
 (36)
0 0 0 0 0
0 1 1 3 2 7 7 19

2. Then, applying the ”qr” command from MatLab in order to decompose the controlla-

7
Karen Reyes ECE 602
FunWork 4 Tuesday, November 05, 2019

bility matrix into an upper triangular form where U = Q ∗ R, therefore T = Q−1

 
1 0 0 0
0 0 −1 0
Q= 0 0
 (37)
0 1
0 −1 0 0
 
−1 1 0 0 00 0 0 −10 −1 0 00
T =Q = (38)
0 1 0

3. Finally, applying the formulas in order to find the pair (A,B) in its news coordinates:
 
−1 −1 −1 0
−1 2 1 1
à = T AT −1 = Q−1 AQ = 
−1 0
 (39)
1 0
0 0 0 4
 
1 1
0 −1
B̃ = T B = Q−1 B =  
0 0  (40)
0 0

4. Including the MatLab script with the used functions and process:

Problem 5
Transform the pairs into their multi-input controller forms.
Pair Number 1:    
0 0 1 0 1 0
1 0 2 0
 0 0
A1 = 
0 and B1 =   (41)
1 3 0 0 0
0 0 21 5 0 1

8
Karen Reyes ECE 602
FunWork 4 Tuesday, November 05, 2019

1. The controllability matrix of the pair (A,B) is found by applying the formula, where
b1 and b2 are the columns of the matrix B:
 
U = b1 Ab1 A 2 b1 b2 (42)
 
1 0 0 0
0 1 0 0
U =0 0
 (43)
1 0
0 0 0 1
2. Then, choosing the vectors q1 and q2 in order to be able to construct the transformation
matrix T , where q1 and q2 are the last rows of the inverse matrix U.
 
q1 = 0 0 1 0 (44)
 
q1 = 0 0 0 1 (45)
3. Calculating the transformation matrix T:
   
q1 0 0 1 0
 q1 A  0 1 3 0
T =
q1 A2  = 1
   (46)
3 11 0
q2 0 0 0 1
4. Finally, applying the formulas in order to find the pair (A,B) in its news coordinates:
 
0 1 0 0
0 0 1 0
à = T AT −1 = 

 1
 (47)
2 3 0
−21 0 0 5
 
0 0
0 0
B̃ = T B = 1
 (48)
0
0 1
5. Including the MatLab script with the used functions and process:

9
Karen Reyes ECE 602
FunWork 4 Tuesday, November 05, 2019

Pair Number 2:    
−1 1 0 1 1 1
1 1 0 0 2 0
A2 = 
0
 and B2 =   (49)
0 4 0 1 0
1 1 −1 2 0 1

1. The controllability matrix of the pair (A,B) is found using the formula in Matlab and
b1 and b2 are the columns of the matrix B:

 
CM = U = b1 Ab1 b2 Ab2 (50)
 
1 1 1 0
2 3 0 1
U =1 4
 (51)
0 0
0 2 1 3

2. Then, extracting the second and fourth rows of the inverse matrix U which will be the
row vectors q1 and q2 :  
q1 = −0.5 −0.15 0.35 0.05 (52)

 
q2 = −0.25 0.25 −0.25 0.25 (53)

3. Calculating T using the following formula:

   
q1 −0.05 −0.15 0.35 0.05
q1 A −0.05 −0.15 1.35 0.05
T =  q2  = −0.25 0.25 −0.25 0.25
   (54)
q2 A 0.75 0.25 −1.25 0.25

4. Finally, applying the formulas in order to find the pair (A,B) in its news coordinates:
 
0 1 0 0
−4 5 0 0
à = T AT −1 =
0
 (55)
0 0 1
3 −3 4 1
 
0 0
1 0
B̃ = T B = 
0
 (56)
0
0 1

10
Karen Reyes ECE 602
FunWork 4 Tuesday, November 05, 2019

5. Including the MatLab script with the used functions and process:

Problem 6
For the pair use its controller form to find the matrix K:
   
0 0 1 0 1 0
1 0 2 0 0 0
A= 0 1 3 and B =   (57)
0 0 0
0 0 21 5 0 1

1. The controller form of this matrix was found in a previous problem, and it is:
   
0 1 0 0 0 0
 0 0 1 0 0
  0
(Ã, B̃) = 
 1 ,  (58)
2 3 0 1 0
−21 0 0 5 0 1

2. Obtaining the characteristic polynomial by multiplying the given eigenvalues.

eig(A − BK) = {−2, −3 + j, −3 − j, −4} (59)

P (s) = s4 + 12s3 + 54s2 + 108s + 80 (60)

3. The realization matrix can be defined as follow, since the characteristic polinomial is
know:  
0 1 0 0
 0 0 1 0 
à − B̃ K̃ = 
 0
 (61)
0 0 1 
−80 −108 −54 −12

11
Karen Reyes ECE 602
FunWork 4 Tuesday, November 05, 2019

4. Calculate B̃ K̃ using the following formula:

B̃ K̃ = Ã − (Ã − B̃ K̃) (62)

   
0 1 0 0 0 1 0 0
 0 0 1 0
− 0 0 1 0 

B̃ K̃ = 
 1
 (63)
2 3 0  0 0 0 1 
−21 0 0 5 −80 −108 −54 −12
 
0 0 0 0
0 0 0 0
B̃ K̃ =   (64)
1 2 3 −1
59 108 54 17
 
1 2 3 −1
K̃ = (65)
59 108 54 17

5. Finally calculating K since K = K̃T , where T is the transformation matrix:

 
3 11 40 −1
K = K̃T = (66)
54 207 977 17

6. The MatLab script is attached:

12
Karen Reyes ECE 602
FunWork 4 Tuesday, November 05, 2019

Problem 7
Transform the pair into its observer form.
 
0 0 1 0  
1 0 2 0
A=  and C = 1 0 0 0
0 1 3 0 0 0 0 1
0 0 21 5

1. In order to compute the observer form of the pair (A,C), first the pair has to be
transformed to (AT , C T ), then follow the same process like the controller form

2. Find the controllability matrix of the pair (AT , C T ) using the following formula, and
MatLab:

 
U = cT1 AT cT1 (AT )2 cT1 cT2 (67)
 
1 0 0 0
0 0 1 0
U =
0 1 3 0
 (68)
0 0 0 1

3. Computing the U inverse U −1 :

 
1 0 0 0
0 −3 1 0
U −1 =
0 1 0 0
 (69)
0 0 0 1

4. Then, extracting the third and fourth row of the inverse matrix U which will be the
row vectors q1 and q2 :  
q1 = 0 1 0 0 (70)

 
q2 = 0 0 0 1 (71)

5. Calculating T using the following formula:

   
q1 0 1 0 05
 q1 AT  0 0 1 0 
T = q1 (AT )2  = 1
   (72)
2 3 −21
q2 0 0 0 1

13
Karen Reyes ECE 602
FunWork 4 Tuesday, November 05, 2019

6. Finally, applying the formulas in order to find the pair (A,B) in its news coordinates:
 
0 0 1 0
1 0 2 0
à = T −T AT T = 

0
 (73)
1 3 0
0 0 −105 5
 
T 0 0 1 0
C̃ = CT = (74)
0 0 −21 1

7. Including the MatLab script with the used functions and process:

Problem 8
For the pair use its observer form to find the matrix L:
 
0 0 1 0  
1 0 2 0
A=  and C = 1 0 0 0 (75)
0 1 3 0 0 0 0 1
0 0 21 5

1. The observer form of this matrix was found in a previous problem, and it is:
 
0 0 1 0  
1 0 2 0 0 0 1 0
(Ã, C̃) =  , (76)
0 1 3 0 0 0 −21 1
0 0 −105 5

14
Karen Reyes ECE 602
FunWork 4 Tuesday, November 05, 2019

2. Obtaining the characteristic polynomial by multiplying the given eigenvalues.

eig(A − LC) = {−2, −3 + j, −3 − j, −4} (77)
P (s) = s4 + 12s3 + 54s2 + 108s + 80 (78)
3. The realization matrix can be defined as follow, since the characteristic polynomial is
know:  
0 0 0 −80
1 0 0 −108
à − L̃C̃ = 
0 1 0 −54 
 (79)
0 0 1 −12

4. Calculate L̃C̃ using the following formula:

 
1681 80
2270 108
L̃ =  
1137 54  (80)
251 17

5. Finally calculating L since L = T T L̃, where T is the transformation matrix:

 
1137 54
 3955 108 
L = T T L̃ = 
 5681
 (81)
270 
−23626 −1117
6. The MatLab script is attached:

15