1
UNIVERSITY OF BUEA REPUBLIC OF CAMEROON
P.O.Box:63 Peace - Work - Fatherland
Buea, South West Region
Cameroon
Tel: (+237) 33322134
Fax: (+237) 33322272
FACULTY OF ENGINEERING AND TECHNOLOGY
DEPARTMENT OF ELECTRICAL AND ELECTRONIC ENGINEERING
MASTERS OF TELECOMMUNICATION AND NETWORK ENGINEERING
STATE SPACE CONTROL TUTORIAL SOLUTION
BATCHANOU TATAP PERETS ARNAUD – FE20P032
Option: Masters 1, Telecommunication & Network Engineering
Course Code: FET653
Course Title: Modern Control of Systems
Course Instructor: Prof. TANYI Emmanuel
University of Buea
Academic Year 2020/2021
2
EXERCISE 1.
Consider the system:
𝑌(𝑆) 𝑆+1
=
𝑈(𝑠) 𝑆 2 + 7𝑆 + 10
A) Analysis.
i. Obtain the controllable state space realization of the system
Using direct programming method.
𝑌(𝑆) 𝑌(𝑆) 𝐸(𝑆) 𝑆 +1
= ∗ = 2
𝑈(𝑠) 𝐸(𝑠) 𝑈(𝑠) 𝑆 + 7𝑆 + 10
Dividing top and bottom by the highest power of S we obtain:
𝑌(𝑆) 𝐸(𝑆) 𝑆 −1 + 𝑆 −2
∗ =
𝐸(𝑠) 𝑈(𝑠) 1 + 7𝑆 −1 + 10𝑆 −2
Therefore
𝑌(𝑆) 𝐸(𝑆) 1
= 𝑆 −1 + 𝑆 −2 𝑎𝑛𝑑 =
𝐸(𝑠) 𝑈(𝑠) 1 + 7𝑆 + 10𝑆 −2
−1
𝑌(𝑆) = 𝐸(𝑠) 𝑆 −1 + 𝐸(𝑠) 𝑆 −2
𝐸(𝑆) = 𝑈(𝑠) − 7𝐸(𝑆) 𝑆 −1 + 10𝐸(𝑆) 𝑆 −2
Analog diagram:
State space equation
𝑥 1̇ = 𝑥 2
𝑥 2̇ = 𝑢 − 7𝑥 2 − 10𝑥 1
𝑦 = 𝑥1 + 𝑥2
3
Vector matrix format;
0 1 𝑥1 0 𝑢
𝑥̇ = [ ] [𝑥 ] + [ ] [𝑢 1 ]
−10 −7 2 1 2
𝑦 = [1 1] 𝑥
ii. Using iterative programming
𝑌(𝑆) 𝑌(𝑆) 𝐸(𝑆) 𝑆+1 𝑆+1 1
= ∗ = = ( ) ( )
𝑈(𝑠) 𝐸(𝑠) 𝑈(𝑠) 𝑆 2 + 7𝑆 + 10 𝑆+ 2 𝑆 +5
𝑆+1 1+𝑆 −1
Applying direct programming to (𝑆+2 ) yields
1+2𝑆 −1
1 1
Applying direct programming to (𝑆+5 ) yields 1+5𝑆 −1
Analog diagram:
State equation;
𝑥 1̇ = 𝑢 − 2𝑥 1
𝑥̇ 2 = 𝑢 − 2𝑥 1 + 𝑥 1 − 5𝑥 2 ⇒ 𝑥̇ 2 = 𝑢 − 𝑥 1 − 5𝑥 2
𝑦 = 𝑥2
Vector Matrix format
−2 0 𝑥 1 1 𝑢
𝑥̇ = [ ] [𝑥 ] + [ ] [𝑢 1 ]
−1 −5 2 1 2
𝑦 = [ 0 1] 𝑥
iii. Transformation from controllable realization in (i) into parallel form.
0 1 𝑥1 0 𝑢1
𝑥̇ = [ ] [𝑥 ] + [ ] [𝑢 ]
−10 −7 2 1 2
𝑦 = [1 1] 𝑥
Eigenvalues
4
|λI − A| = 0
0 1
|[λ 0] − [ ]| = 0
0 λ −10 −7
λ −1
|[ ]| = 0.
10 (λ + 7)
λ2 + 7λ + 10 = 0.
λ1 = −2 𝑎𝑛𝑑 λ2 = −5
Eigenvector associated to λ1 𝑎𝑛𝑑 λ2
𝑤
For λ1 = −2. Let 𝑊1 = [𝑤1 ]
2
[λ1 𝐼 − 𝐴]𝑊1 = 0
−2 0 0 1 𝑤
[[ ]−[ ]]*[𝑤1 ] = 0.
0 −2 −10 −7 2
−2 −1 𝑤1
[ ][ ] = 0
10 5 𝑤2
−2𝑤1 − 𝑤2 = 0
10𝑤1 + 5𝑤2 = 0
1
For 𝑤1 = 1 𝑤2 = −2. 𝑡ℎ𝑒𝑟𝑒𝑓𝑜𝑟𝑒 𝑊1 = [ ]
−2
For λ2 = −5
[λ2 𝐼 − 𝐴]𝑊2 = 0
−5 0 0 1 𝑤1
[[ ]−[ ]] [𝑤 ] = 0.
0 −5 −10 −7 2
−5 −1 𝑤1
[[ ]] [𝑤 ] = 0
10 2 2
−5𝑤1 − 𝑤2 = 0
10𝑤1 + 2𝑤2 = 0)
1
For 𝑤1 = 1 𝑤2 = −5 𝑡ℎ𝑒𝑟𝑒𝑓𝑜𝑟𝑒 𝑊1 = [ ]
−5
Modal matrix is;
𝑊 = (𝑤1 𝑤2 ) = [ 1 1
]
−2 −5
5
1 1
[−5 −1]. 𝑊 −1 = − [−5 −1].
𝑊 −1 =
−5 + 2 2 1 3 2 1
Parallel form.
𝐹 = 𝑊 −1 𝑊𝐴
1 0 1 1 1
𝐹 = − [−5 −1] [ ][ ]
3 2 1 −10 −7 −2 −5
1
𝐹 = − [−5 −1] [−2 −5]
3 2 1 4 25
1 6 0
𝐹=− [ ]
3 0 15
−2 0
𝐹 =[ ]
0 −5
𝑔 = 𝑊 −1 𝑏
1 −5 −1 0
𝑔=− [ ][ ]
3 2 1 1
1 −1
𝑔=− [ ]
3 1
1
𝑔=[ 3 ];
−1
3
ℎ𝑇 = 𝐶 𝑇 𝑊
1 1
ℎ𝑇 = [ 1 1] [ ]
−2 −5
ℎ𝑇 = [−1 −4]
Parallel realization is therefore;
𝑍̇ = 𝐹𝑧 + 𝑔𝑢 𝑦 = ℎ𝑇 𝑧
1
−2 0
𝑍̇ = [ ] 𝑧 + [ 3 ]𝑢 𝑦 = [−1 −4]𝑧
0 −5 −1
3
Where 𝑥 = 𝑊𝑧
iv. Transform the iterative programming realization into parallel form.
−2 0 𝑥1 1 𝑢
𝑥̇ = [ ] [𝑥 ] + [ ] [𝑢 1 ]
−1 −5 2 1 2
6
𝑦 = [0 1] 𝑥
Eigenvalues
|λI − A| = 0
λ 0 −2 0
|[ ]−[ ]| = 0
0 λ −1 −5
(λ − 2) 0
|[ ]| = 0.
1 (λ + 5)
λ2 + 7λ + 10 = 0.
λ1 = −2 𝑎𝑛𝑑 λ2 = −5
Eigenvector associated to λ1 𝑎𝑛𝑑 λ2
𝑤1
For λ1 = −2. Let 𝑊1 = [𝑤 ]
2
[λ1 𝐼 − 𝐴]𝑊1 = 0
−2 0 −2 0 𝑤
[[ ]−[ ]] [𝑤1 ] = 0.
0 −2 −1 −5 2
0 0 𝑤
[[ ]]*[𝑤1 ] = 0
1 3 2
0𝑤1 − 0𝑤2 = 0
𝑤1 + 3𝑤2 = 0
−3
for 𝑤2 = 1; 𝑤1 = −3. 𝑡ℎ𝑒𝑟𝑒𝑓𝑜𝑟𝑒 𝑊1 = [ ]
1
For λ2 = −5
[λ2 𝐼 − 𝐴]𝑊1 = 0
−2 0 𝑤
[[−5 0 ] − [ ]] [𝑤1 ] = 0
0 −5 −1 −5 2
−3 0 𝑤
[[ ]]*[𝑤1 ] = 0
1 0 2
−3𝑤1 + 0𝑤2 = 0)
𝑤1 + 0𝑤2 = 0
𝑤1 = 0 𝑤2 ∈ 𝑅 , 𝑙𝑒𝑡 𝑤2 = 1
0
𝑊2 = [ ]
1
Modal matrix
𝑊 = (𝑤1 𝑤2 ) = [−3 0
]
1 1
7
1 1 0 1 1 0
𝑊 −1 = [ ]. 𝑊 −1 = − [ ].
−3 + 0 −1 −3 3 −1 −3
Parallel form.
𝐹 = 𝑊 −1 𝑊𝐴
1 1 0 −2 0 −3 0
𝐹 =− [ ][ ][ ]
3 −1 3 −1 −5 1 1
1 1 0 6 0
𝐹 =− [ ][ ]
3 −1 −3 −2 −5
1 6 0 −2 0
𝐹 =− [ ] ; 𝐹=[ ]
3 0 15 0 −5
𝑔 = 𝑊 −1 𝑏
1 1 0 1
𝑔=− [ ][ ]
3 −1 −3 1
1 1
𝑔=− [ ] ;
3 −4
1
−
𝑔 = [ 3]
4
3
ℎ𝑇 = 𝐶 𝑇 𝑊
−3 0
ℎ𝑇 = [0 1] [ ]
1 1
ℎ𝑇 = [1 1]
Parallel realization is therefore;
𝑍̇ = 𝐹𝑧 + 𝑔𝑢 𝑦 = ℎ𝑇 𝑧
1
−
−2 0
𝑍̇ = [ ] 𝑧 + [ 3] 𝑢 𝑦 = [1 1] 𝑧
0 −5 4
3
v. Conclusion
From iii and iv above we obtain the same F matrix but not g and ℎ𝑇 . this means that
vi. Using controllable realization of (i)
Controllability of the system
8
𝑃 = [𝑏 𝐴𝑏]
0 1 0
𝐴𝑏 = − [ ][ ]
−10 −7 1
1
𝐴𝑏 = [ ]
−7
0 1
𝑃=[ ]
1 −7
|𝑃| = −1
None zero value implies rank is 2
Rank of system equals n therefore the system is controllable.
Observability of the system
𝑇
𝑄 = [ 𝐶 𝑇]
𝐴𝐶
0 1
𝐴𝐶 𝑇 = − [ ] [1 1]
−10 −7
𝐴𝐶 𝑇 = [−10 −6]
1 1
𝑄=[ ]
−10 −6
|𝑄| = 4
None zero value implies rank equals 2. Also rank equals n therefore the system is
observable.
vii. Iterative programming realization (ii)
Controllability of the system
𝑃 = [𝑏 𝐴𝑏]
−2 0 1
𝐴𝑏 = − [ ][ ]
−1 −5 1
−2
𝐴𝑏 = [ ]
−6
1 −2
𝑃=[ ]
1 −6
|𝑃| = −4
None zero value implies the rank equals 2. Also, rank equals n hence, the system is
controllable.
Observability of the system
𝑇
𝑄 = [ 𝐶 𝑇]
𝐴𝐶
9
−2 0
𝐴𝐶 𝑇 = [0 1] [ ]
−1 −5
𝐴𝐶 𝑇 = [−1 −5]
0 1
𝑄=[ ]
−1 −5
|𝑄| = −1
Non zero determinant implies rank equals 2. Also rank equals n therefor the system is
observable.
viii. Conclusion draw from (vi) and (vii).
The system is controllable and observable irrespective of the programming used.
MATLAB Code Used
i) X(t)
A = [0, 1; -10, -7]
b = [0; 1]
C= [ 1, 1]
X0 = [0; 1]
eig(A)
[W, D] = eig(A)
syms t
x(t) = (W * diag(exp(diag(D*t))) *inv(W) * X0)
display(x(t))
ii) State Transition Matrix Φ(t)
A = [0, 1; -10, -7]
b = [0; 1]
C= [ 1, 1]
X0 = [0; 1]
I = exp(A)
syms t
phi = exp(A*t)
display phi
phi = [ 1, exp(t)]
[ exp(-10*t), exp(-7*t)]
10
𝟎
iii) X(t) when X0 = [ ] and U = 1
𝟏
A = [0, 1; -10, -7]
b = [0; 1]
C= [ 1, 1]
X0 = [0; 1]
I = exp(A)
syms t
phi = exp(A*t)
x(t) = phi*X0 + b
display x(t)
EXERCISE 2.
Consider the system:
𝑌(𝑆) 𝑆+5
=
𝑈(𝑠) 𝑆3 + 6𝑆 2 + 11𝑆 + 6
A) Analysis.
i. Obtain the controllable state space realization of the system
Using direct programming method.
𝑌(𝑆) 𝑌(𝑆) 𝐸(𝑆) 𝑆 +5
= ∗ = 3
𝑈(𝑠) 𝐸(𝑠) 𝑈(𝑠) 𝑆 + 6𝑆 2 + 11𝑆 + 6
Dividing top and bottom by highest power of s, we obtain
𝑌(𝑆) 𝐸(𝑆) 𝑆 −2 + 5𝑆 −3
∗ =
𝐸(𝑠) 𝑈(𝑠) 1 + 6𝑆 −1 + 11𝑆 −2 + 6𝑆 −3
𝑌(𝑆) 𝐸(𝑆) 1
= 𝑆 −2 + 5𝑆 −3 𝑎𝑛𝑑 =
𝐸(𝑠) 𝑈(𝑠) 1 + 6𝑆 −1 + 11𝑆 −2 + 6𝑆 −3
𝑌(𝑆) = 𝐸(𝑠) 𝑆 −2 + 5𝐸(𝑠) 𝑆 −3
𝐸(𝑆) = 𝑈(𝑠) − 6𝐸(𝑆) 𝑆 −1 + 11𝐸(𝑆) 𝑆 −2 + 6𝐸(𝑆) 𝑆 −3
11
Analog diagram:
State space equation;
𝑥 1̇ = 𝑥 2
𝑥 2̇ = 𝑥 3
𝑥 3̇ = 𝑢 − 6𝑥1 − 11𝑥 2 − 6𝑥 3
𝑦 = 5𝑥 1 + 𝑥 2
Vector matrix format;
0 1 0 0
𝑥̇ = ⌈ 0 0 1 ⌉ 𝑥 + [ 0] 𝑢
−6 −11 −6 1
𝑦 = [5 1 0] 𝑥
ii. Using iterative programming to obtain state space realization
𝑌(𝑆) 𝑌(𝑆) 𝐸(𝑆) 𝑆 +5 𝑆+5 1 1
= ∗ = =( )( )( )
𝑈(𝑠) 𝐸(𝑠) 𝑈(𝑠) 𝑆3 2
+ 6𝑆 + 11𝑆 + 6 𝑆+1 𝑆+2 𝑆+3
𝑆+5 1+5𝑆 −1
Applying direct programming to ( ) we obtain
𝑆+1 1+𝑆 −1
1 𝑆 −1
Applying direct programming to (𝑆+2 ) we obtain 1+2𝑆 −1
1 𝑆 −1
Applying direct programming to (𝑆+3 ) we obtain 1+3𝑆 −1
12
Analog representation;
State space equation;
𝑥 1̇ = 𝑢 − 𝑥 1
𝑥 2̇ = 𝑢 + 4𝑥 1 −2𝑥 2
𝑥 3̇ = 𝑢 + 4𝑥 1 −2𝑥2 + 𝑥 2 − 3𝑥 3
𝑥 3̇ = 𝑥 2 − 3𝑥 3
𝑦 = 𝑥3
Vector matrix format;
−1 0 0 1
𝑥̇ = ⌈ 4 −2 0 ⌉ 𝑥 + [1] 𝑢
0 1 −3 0
𝑦 = [0 0 1] 𝑥
iii. Determine the controllability and observability using control realization in (i).
Controllability
𝑃 = [𝑏 𝐴𝑏 𝐴2 𝑏]
0 1 0 0
𝐴𝑏 = ⌈ 0 0 1 ⌉ [0]
−6 −11 −6 1
0
𝐴𝑏 = [ 1 ]
−6
0 1 0 0
2 ⌈
𝐴 𝑏= 0 0 1 ⌉[ 1 ]
−6 −11 −6 −6
1
𝐴2 𝑏 = [−6]
25
13
0 0 1
𝑃 = ⌈0 1 −6⌉
1 −6 25
|𝑃| = 0[25 + 36] − 0[0 − 6] + 1[0 − 1]
|𝑃| = −1
Non zero determinant implies rank equals 3. Also rank rank equals 3 hence system is
controllable.
Observability
𝐶𝑇
𝑄 = [ 𝐶𝑇𝐴 ]
𝐶 𝑇 𝐴2
0 1 0
𝐴𝐶 𝑇 = [5 1 0] ⌈ 0 0 1 ⌉ = [0 5 1]
−6 −11 −6
0 1 0
2 𝑇
𝐴 𝐶 = [0 5 1] ⌈ 0 0 1 ⌉ = [−6 −11 −1]
−6 −11 −6
5 1 0
𝑄 = ⌈0 5 1⌉
6 −11 −1
|𝑄| = 0(9 + 15) − 0(0 − 12) + 1(0 − 4) = −4
Non zero determinant implies rank equals 3. Also, n equals 3 hence system is observable.
iv. Determine the controllability and observability using iterative programming
realization in.
Controllability
𝑃 = [𝑏 𝐴𝑏 𝐴2 𝑏]
−1 0 0 1
𝐴𝑏 = ⌈ 4 −2 1 ⌉ [1]
0 1 −3 0
−1
𝐴𝑏 = [ 2 ]
1
−1 0 0 −1
𝐴2 𝑏 = ⌈ 4 −2 0 ⌉[ 2 ]
0 1 −3 1
14
1
𝐴2 𝑏 = [−8]
−1
1 −1 1
𝑃 = ⌈1 2 −8⌉
0 1 −1
|𝑃| = 1[−2 + 8] + 1[−1 − 0] + 1[1 − 0]
|𝑃| = 6
Non zero determinant implies rank equals 3. Also rank rank equals 3 hence system is
controllable.
Observability
𝐶𝑇
𝑄 = [ 𝐶𝑇𝐴 ]
𝐶 𝑇 𝐴2
−1 0 0
𝑇
𝐶 𝐴 = [0 0 1] ⌈ 4 −2 1 ⌉ = [0 1 − 3]
0 1 −3
−1 0 0
2 𝑇
𝐴 𝐶 = [0 1 ]
−3 4⌈ −2 1 ⌉ = [4 − 5 10]
0 1 −3
0 0 1
𝑄 = ⌈0 1 −3⌉
4 −5 10
|𝑄| = 0[10 − 15] − 0[0 − 12] + 1[0 − 4] = −4
Non zero determinant implies rank equals 3. Also n=3 hence the matrix is observable.
v. What conclusion can you draw from (iii) and (iv).
The controllability and observability of the system remains the same irrespective of the
programming method used.
15
MATLAB Code Used
i) State Transition Matrix Φ(t)
A = [0, 1 0; 0, 0, 1; -6, -11, -6]
b = [0;0;1]
C= [ 5, 1,0]
X0 = [0]
I = exp(A)
syms t
phi = exp(A*t)
display phi
phi = [ 1, exp(t)]
[ exp(-10*t), exp(-7*t)]
𝟎
ii) X(t) when X0 = [ ] and U = 1
𝟏
A = [0, 1; -10, -7]
b = [0; 1]
C= [ 1, 1]
X0 = [0; 1]
I = exp(A)
syms t
phi = exp(A*t)
x(t) = phi*X0 + b
display x(t)
