EET2415: CONTROL SYSTEMS II

PRESENTED BY

Dr. Linus A. Aloo, PhD.

STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo

 1.0 STATE VARIABLE METHODS
1. 5. State Equations from Transfer Functions
Consider a system defined by

y (n) + a1 y (n−1) + ⋯ + a n−1 yሶ + a n y = b0 un + b1 un−1 + ⋯ + bn−1 uሶ + bn u (1.17)

where 𝑢 is the input and 𝑦 is the output.

Assuming zero initial conditions, eqn, (1.18) becomes:

𝑌(𝑠) 𝑏0 𝑠 𝑛 + 𝑏1 𝑠 𝑛−1 + ⋯ + 𝑏𝑛−1 𝑠 + 𝑏𝑛

𝑇 𝑠 = = 𝑛 (1.18)
𝑈(𝑠) 𝑠 + 𝑎1 𝑠 𝑛−1 + ⋯ + 𝑎𝑛−1 𝑠 + 𝑎𝑛

The state space representation of the system defined by equation (1.17) and (1.18) can be presented
in controllable canonical form, observable canonical form and diagonal canonical form

STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo

 1.0 STATE VARIABLE METHODS
1. 5. State Equations from Transfer Functions
Controllable canonical form
The controllable canonical form (CCF) of the state space representation is given by

𝑥ሶ1 0 1 0 . . . 0 𝑥1 0
𝑥ሶ 2 0 0 1 . . . 0 𝑥2 0
. . . . . . .
= . + 𝑢 (1.19)
. . . . . .
𝑥ሶ 𝑛−1 0 0 0 . . . 1 𝑥𝑛−1 0
𝑥ሶ 𝑛 −𝑎𝑛 −𝑎𝑛−1 −𝑎𝑛−2 . . . −𝑎1 𝑥𝑛 1
and 𝑥1
𝑥2
𝑦 = 𝑏𝑛 − 𝑎𝑛 𝑏0 𝑏𝑛−1 − 𝑎𝑛−1 𝑏0 . . . 𝑏1 − 𝑎1 𝑏0 . + 𝑏0 𝑢 (1.20)
.
𝑥𝑛
 The controllable canonical form is useful in the design of control systems using the pole-placement
approach.

STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo

 1.0 STATE VARIABLE METHODS
1. 5. State Equations from Transfer Functions
Observable canonical form
 For the same system whose transfer function is given by eqn. 1.18, the observable canonical form
(OCF) of the state space representation is given by

𝑥ሶ1 0 0 . . . 0 −𝑎𝑛 𝑥1 𝑏𝑛 − 𝑎𝑛 𝑏0
𝑥ሶ 2 1 0 . . . 0 −𝑎𝑛−1 𝑥2 𝑏𝑛−1 − 𝑎𝑛−1 𝑏0
. . . . . . .
= . + 𝑢 (1.21)
. . . . . .
𝑥ሶ 𝑛−1 0 0 . . . 0 −𝑎2 𝑥𝑛−1 𝑏2 − 𝑎2 𝑏0
𝑥ሶ 𝑛 0 0 . . . 1 −𝑎1 𝑥𝑛 𝑏1 − 𝑎1 𝑏0

and 𝑥1
𝑥2
.
𝑦= 0 0 . . . 0 1 . + 𝑏0 𝑢 (1.22)
𝑥𝑛−1
𝑥𝑛
STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo
 1.0 STATE VARIABLE METHODS
1. 5. State Equations from Transfer Functions
Diagonal canonical form
 This form is also known as canonical variable form or normal form. The system matrix A in this case is obtained as a diagonal matrix.
 If the denominator polynomial has distinct roots then the transfer function can be written as

𝑌(𝑠) 𝑏0 𝑠 𝑛 + 𝑏1 𝑠 𝑛−1 + . . . +𝑏𝑛−1 𝑠 + 𝑏𝑛

= (1.23)
𝑈(𝑠) 𝑠 + 𝑝1 𝑠 + 𝑝2 . . . (𝑠 + 𝑝𝑛−1)(𝑠 + 𝑝𝑛 )

𝑐1 𝑐2 𝑐𝑛−1 𝑐𝑛
= 𝑏0 + + ...+ + (1.24)
(𝑠 + 𝑝1 ) (𝑠 + 𝑝2 ) (𝑠 + 𝑝𝑛−1 ) (𝑠 + 𝑝𝑛 )
 The diagonal canonical form of the state space representation of the system is given by

𝑥ሶ 1 −𝑝1 0 . . . 0 0 𝑥1 1
𝑥ሶ 2 0 −𝑝1 . . . 0 0 𝑥2 1
. . . . . . .
= . + 𝑢 (1.25)
. . . . . .
𝑥ሶ 𝑛−1 0 0 . . . −𝑝𝑛−1 0 𝑥𝑛−1 1
𝑥ሶ 𝑛 0 0 . . . 0 −𝑝𝑛 𝑥 𝑛 1
and 𝑥1
𝑥2
.
𝑦 = 𝑐1 𝑐2 . . . 𝑐𝑛−1 𝑐𝑛
. + 𝑏0 𝑢 (1.26)
𝑥𝑛−1
𝑥𝑛
STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo
 1.0 STATE VARIABLE METHODS
1. 5. State Equations from Transfer Functions
Example 1.4: Obtain the state space representation of the following system in controllable canonical form
(CCF).
𝑦ഺ + 2𝑦ሷ + 3𝑦ሶ + 6y = 𝑢ሷ − 𝑢ሶ + 2u
Solution
In the given differential equation,
𝑎1 = 2, 𝑎2 = 3, 𝑎3 = 6 and 𝑏0 = 0, 𝑏1 = 1, 𝑏2 = −1
and 𝑏3 = 2. Subtituting in eqns. (1.30) and (1.31), we get

𝑥ሶ 1 0 1 0 𝑥1 0
𝑥ሶ 2 = 0 0 1 𝑥2 + 0 𝑢
𝑥ሶ 3 −6 −3 −2 𝑥3 1

𝑥1
y= 2 −1 1 𝑥2
𝑥3

STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo

 1.0 STATE VARIABLE METHODS
1. 5. State Equations from Transfer Functions
Example 1.5: Consider the system given by
𝑌(𝑠) 𝑠+3
=
𝑈(𝑠) 𝑠 2 + 3𝑠 + 2
Obtain the state space representation of the system in:
i. Controllable Canonical Form
ii. Observable Canonical Form
iii. Diagonal Canonical Form
Solution
𝑛 = 2 𝑏0 = 0 𝑏1 = 1 𝑏2 = 3 𝑎1 = 3 𝑎2 = 2
Controllable canonical form of the state space representation for the given system
𝑥ሶ1 0 1 𝑥1 0 0 1 𝑥1 0
= + 𝑢 = + 𝑢
𝑥ሶ 2 −𝑎2 −𝑎1 𝑥2 1 −2 −3 𝑥2 1

𝑥
𝑦 = 𝑏2 − 𝑎2 𝑏0 𝑏1 − 𝑎1 𝑏0 𝑥1 + 𝑏0 𝑢
2
𝑥1
= 3 1 𝑥
2

STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo

 1.0 STATE VARIABLE METHODS
1. 5.2. State Equations from Transfer Functions
Example 1.5:
ii) Observable canonical form of the state space representation for the given
system
𝑥ሶ 1 0 −𝑎2 𝑥1 𝑏2 − 𝑎2 𝑏0
= + 𝑢
𝑥ሶ 2 1 −𝑎1 𝑥2 𝑏1 − 𝑎1 𝑏0

0 −2 𝑥1 3
= 𝑥 + 𝑢
1 −3 2 1
𝑥1
𝑦 = 0 1 𝑥 + 𝑏0 𝑢
2

𝑥1
= 0 1 𝑥
2

STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo

 1.0 STATE VARIABLE METHODS
1. 5.2. State Equations from Transfer Functions
Example 1.5:
iii) The diagonal canonical form of the state space representation for the given system

𝑠+3 𝑠+3 𝑐1 𝑐2
= = +
𝑠 2 + 3𝑠 + 2 (𝑠 + 1)(𝑠 + 2) (𝑠 + 1) (𝑠 + 2)

𝑏0 = 0 𝑐1 = 2 𝑐2 = −1 𝑝1 = 1 𝑝2 = 2

𝑥ሶ 1 −𝑝1 0 𝑥1 1
= + 𝑢
𝑥ሶ 2 0 −𝑝2 𝑥2 1

−1 0 𝑥1 1
= + 𝑢
0 −2 𝑥2 1

𝑥1
𝑦 = 𝑐1 𝑐2
𝑥2 + 𝑏0 𝑢
𝑥1
= 2 −1 𝑥
2

STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo

 1.0 STATE VARIABLE METHODS
1.6. Laplace Transform Technique
Recall
1. The Laplace transform of a function 𝑓(𝑡), denoted by ℒ 𝑓(𝑡) = 𝐹 𝑠 , is defined as the integral
∞
ℒ 𝑓(𝑡) = 𝐹 𝑠 = ‫׬‬0 𝑒 −𝑠𝑡 𝑓 𝑡 𝑑𝑡 (1.27)
2. The Laplace transform for first order and second order systems require specification of initial
conditions and are respectively defined in Eqns. (1.28a and 1.28b)

ℒ 𝑓 ′ (𝑡) = 𝑠ℒ 𝑓(𝑡) − 𝑓 0 (1.28a)

ℒ 𝑓 ′′ (𝑡) = 𝑠 2 ℒ 𝑓(𝑡) − 𝑠𝑓 0 − 𝑓 ′ 𝑜 (1.28b)

and so on for Laplace transform of higher derivatives.

Example 1.6
Obtain the Laplace transform of 𝑓(𝑡) = 𝑒 𝑎𝑡
Solution
∞
1
ℒ 𝑓 𝑡 =න 𝑒 −𝑠𝑡 𝑒 𝑎𝑡 𝑑𝑡 =
0 𝑠−𝑎

STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo

 1.0 STATE VARIABLE METHODS
1.6. Laplace Transform Technique

Table 1.1 shows some common Laplace transforms.

Table 1.1: Selected Laplace Transforms
𝐿−1 𝐹 𝑠 = 𝑓(𝑡)
𝐹 𝑠 𝑓(𝑡)
𝑎 𝑎
𝑠

𝑛!
𝑠 𝑛+1 𝑡 𝑛 , 𝑛 = 1,2,3 … . .

1 𝑒 𝑎𝑡
𝑠−𝑎

𝑎 sin 𝑎𝑡
𝑠 2 + 𝑎2
𝑠 cos 𝑎𝑡
𝑠2 + 𝑎2
𝑎 sinh 𝑎𝑡
𝑠2 − 𝑎2
𝑠 cosh 𝑎𝑡
𝑠2 − 𝑎2

STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo

 1.0 STATE VARIABLE METHODS
1.7 Solution of State Equation
1.7.1 Solution of Homogenous Scalar Equation
 Consider the homogenous case

𝑥ሶ 𝑡 = 𝑎𝑥 𝑡 (1.29)

Equation (1.29) can be solved using the Laplace transform method or the power series
method.
Solution by Laplace Transform Method
Using Laplace transform of the homogeneous equation (1.29)

1
𝑠𝑋 𝑠 − 𝑋 0 = 𝑎𝑋(𝑠) → 𝑋 𝑠 = X(0)
𝑠−𝑎

 Taking the inverse Laplace transform 𝑥 𝑡 = 𝑒 𝑎𝑡 𝑥(0)

STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo

 1.0 STATE VARIABLE METHODS
1.7 Solution of State Equation
1.7.2. State Transition Matrix
 The approach to the solution of homogeneous scalar equation can be extended to the solution of
homogeneous state equation:
𝑥ሶ 𝑡 = 𝐴𝑥 𝑡 (1.30)
 Taking Laplace transform

𝑠𝑋 𝑠 − 𝑋 0 = 𝐴𝑋(𝑠)

𝑋 𝑠 = 𝑠𝐼 − 𝐴 −1 𝑋(0)

 Taking the inverse Laplace transform

𝑥 𝑡 = ℒ −1 𝑠𝐼 − 𝐴 −1 𝑥 0 (1.31)

𝑥 𝑡 = 𝑒 𝐴𝑡 𝑥 0 (1.32)

Where 𝑒 𝐴𝑡 = ℒ −1 𝑠𝐼 − 𝐴 −1

𝑒 𝐀𝑡 is called the state transition matrix and contains all information about the free motions of the
system described by (1.30).

STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo

 1.0 STATE VARIABLE METHODS
1.7 Solution of State Equation

1.7.3. Properties of State Transition Matrix

For the time-invariant system,
𝑥ሶ 𝑡 = 𝐀𝑥 𝑡

for which the state transition, matrix 𝚽 t is defined by

𝚽 t = 𝑒 𝐀𝑡 = ℒ −1 𝑠𝐈 − 𝐀 −1

we have the following important properties of 𝚽 t :

1. 𝚽 0 = 𝑒 𝐀0 = 𝐈
2. 𝚽 t = 𝑒 𝐀𝑡 = (𝑒 −𝐴𝑡 )−1 = [(𝚽 −t ]−1 or 𝚽 −1 t = 𝚽 −t

STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo

 1.0 STATE VARIABLE METHODS
1.7 Solution of State Equation
Solution of Non Homogeneous State Equations
 Consider the non-homogeneous state equation described by

xሶ t = 𝐀x t + 𝐁u t (1.33)
where : 𝐱 = 𝑛 −vector,𝐮 = 𝑟 −vector, 𝐀 = 𝑛 × 𝑛 −constant matrix, 𝐁 = 𝑛 × 𝑟 −constant matrix

 It can be shown that the solution to Eqn. (1.33) is:

𝑡 𝑡
𝑥 𝑡 = e𝐀t 𝑥 0 + e𝐀t ‫׬‬0 𝑒 −𝐴𝜏 𝐵𝑢 𝜏 𝑑𝜏 = e𝐀t 𝑥 0 + ‫׬‬0 𝑒 𝐴(𝑡−𝜏) 𝐵𝑢 𝜏 𝑑𝜏 (1.34a)
which can also be written as

𝑡
𝑥 𝑡 = 𝚽 t 𝑥 0 + න 𝚽 (𝑡 − 𝜏) 𝐵𝑢 𝜏 𝑑𝜏 (1.34b)
0
where 𝚽 t = 𝑒 𝐴𝑡

STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo

 1.0 STATE VARIABLE METHODS
1.7 Solution of State Equation

Solution of Non Homogeneous State Equations

 From equation (1.34b), he first term on the right hand side is the response to the initial
conditions and the second term is the response to the input 𝒖 𝒕 .

 Extending the same approach to the solution of homogeneous scalar equation yields

𝑡
𝑥 𝑡 = 𝑒 𝑎𝑡 𝑥 0 + ‫׬‬0 𝑒 𝑎(𝑡−𝜏) 𝐵𝑢 𝜏 𝑑𝜏 (1.35)

 The solution of 𝑥(𝑡) is the sum of a term consisting of the transition of the initial state and a
term arising from the input vector.

STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo

 1.0 STATE VARIABLE METHODS
1.7. Solution of State Equation
Laplace Transform Approach to the Solution of Nonhomogeneous State Equations
 The solution of the nonhomogeneous state equation
𝑥ሶ 𝑡 = 𝐴𝑥 𝑡 + Bu(t) (1.36)
can also be obtained by the Laplace transform approach.
 Taking Laplace transform of this last equation yields

𝑠𝑋 𝑠 − 𝑋 0 = 𝐴𝑋 𝑠 + 𝐵𝑈(𝑠) → 𝑠𝐼 − 𝐴 𝑋 𝑠 = 𝑋 0 + 𝐵𝑈(𝑠)

−1
 Pre-multiplying both sides of this last equation by 𝑠𝐼 − 𝐴 , gives

−1 −1
𝑋 𝑠 = 𝑠𝐼 − 𝐴 𝑋 0 + 𝑠𝐼 − 𝐴 𝐵𝑈(𝑠) → 𝑋 𝑠 = ℒ[𝑒 𝐴𝑡 ]𝑋 0 + ℒ[𝑒 𝐴𝑡 ]𝐵𝑈(𝑠)
 Taking the inverse Laplace transform of the last equation using the convolution integral results in
𝑡
𝑥 𝑡 = 𝑒 𝐴𝑡 𝑥 0 + ‫׬‬0 𝑒 𝐴(𝑡−𝜏) 𝐵𝑢 𝜏 𝑑𝜏 (1.37)

STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo

 1.0 STATE VARIABLE METHODS
1.7 Solution of State Equation
Example 1.7
𝑥ሶ1 0 1 𝑥1 0
= + 𝑢
𝑥ሶ 2 −2 −3 𝑥2 1

(i) Assuming 𝑢 𝑡 = 0, that’s, the system is homogeneous, obtain the state transition matrix Φ(t) and
also the inverse of the state-transition matrix 𝚽−1 t
(ii) Assuming the system is non-homogeneous, i.e. that 𝑢(𝑡) is a unit step function occurring at 𝑡 = 0, or
𝑢 𝑡 = 1(𝑡), obtain the time response of the following system.
Solution
For the given system, the state equation is
0 1
𝐀=
−2 −3
The state transition matrix 𝚽 t is given by

𝚽 t = 𝑒 𝐴𝑡 = ℒ −1 [ 𝑠𝐈 − 𝐀 −1 ]

Since
𝑠 0 0 1 𝑠 −1
𝑠𝐈 − 𝐀 = − =
0 𝑠 −2 −3 2 𝑠+3

STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo

 1.0 STATE VARIABLE METHODS
1.7 Solution of State Equation
Example 1.7: Solution Cont’
the inverse of 𝑠𝐈 − 𝐀 is given by
−1
1 𝑠+3 1 1 𝑠+3 1
𝑠𝐼 − 𝐴 = =
𝑠 3 + 3𝑠 + 2 −2 𝑠 (𝑠 + 1)(𝑠 + 2) −2 𝑠

𝑠+3 1 2 1 1 1
(𝑠 + 1)(𝑠 + 2) (𝑠 + 1)(𝑠 + 2) − −
= = 𝑠+1 𝑠+2 𝑠+1 𝑠+2
−2 𝑠 −2 2 −1 2
+ +
(𝑠 + 1)(𝑠 + 2) (𝑠 + 1)(𝑠 + 2) 𝑠+1 𝑠+2 𝑠+1 𝑠+2
2 1 1 1
− −
𝑠+1 𝑠+2 𝑠+1 𝑠+2
Hence, 𝚽 t = 𝑒 𝐴𝑡 = ℒ −1 [ 𝑠𝐈 − 𝐀 −1 ] = ℒ −1
−2 2 −1 2
+ +
𝑠+1 𝑠+2 𝑠+1 𝑠+2

2𝑒 −𝑡 − 𝑒 −2𝑡 𝑒 −𝑡 − 𝑒 −2𝑡
=
−2𝑒 −𝑡 + 2𝑒 −2𝑡 −𝑒 −𝑡 + 2𝑒 −2𝑡
Noting that 𝚽−1 t = 𝚽 −t = 𝑒 −𝐴𝑡 , we obtain the inverse of the state-transition matrix as
follows:
2𝑒 𝑡 − 𝑒 2𝑡 𝑒 𝑡 − 𝑒 2𝑡
𝚽 −1 t = 𝚽 −t = 𝑒 −𝐴𝑡 =
−2𝑒 𝑡 + 2𝑒 2𝑡 −𝑒 𝑡 + 2𝑒 2𝑡
STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo
 1.0 STATE VARIABLE METHODS
1.8 Solution of State Equation
Example 1.7: Solution Cont’
ii) For this system,
0 1 0
𝐴= ,𝐵 =
−2 −3 1
 This problem can be solved using the Convolution Integral Method or Laplace transform method.
We shall use the Convolution Integral Method , though it’s longer than the other.
 The response to the unit-step input is obtained using the relationship:

𝑡
𝑥 𝑡 = 𝑒 𝐴𝑡 𝑥 0 + ‫׬‬0 𝑒 𝐴(𝑡−𝜏) 𝐵𝑢 𝜏 𝑑𝜏

 The state-transition matrix 𝚽 t = 𝑒 𝐴𝑡 was obtained in part (i) as

2𝑒 −𝑡 − 𝑒 −2𝑡 𝑒 −𝑡 − 𝑒 −2𝑡
𝚽 t = 𝑒 𝐴𝑡 =
−2𝑒 −𝑡 + 2𝑒 −2𝑡 −𝑒 −𝑡 + 2𝑒 −2𝑡

𝚽 t−𝜏 = 𝑒 𝐴(𝑡−𝜏) = 2𝑒 −(𝑡−𝜏) − 𝑒 −2(𝑡−𝜏) 𝑒 −(𝑡−𝜏) − 2𝑒 −2(𝑡−𝜏)

−2𝑒 −(𝑡−𝜏) + 2𝑒 −2(𝑡−𝜏) −𝑒 −(𝑡−𝜏) + 2𝑒 −2(𝑡−𝜏)
STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo
 1.0 STATE VARIABLE METHODS
1.7 Solution of State Equation

Example 1.7: (ii) Solution Cont’

0 𝑒 −(𝑡−𝜏) − 2𝑒 −2(𝑡−𝜏)
𝑒 𝐴 𝑡−𝜏 𝐵𝑢 𝜏 = 𝑒 𝐴(𝑡−𝜏) [1] =
1 −𝑒 −(𝑡−𝜏) + 2𝑒 −2(𝑡−𝜏)
1 1
𝑡 𝑡 𝑒 −(𝑡−𝜏) − 2𝑒 −2(𝑡−𝜏) − 𝑒 −𝑡 + 𝑒 −2𝑡
‫׬‬0 𝑒 𝐴(𝑡−𝜏) 𝐵𝑢 𝜏 𝑑𝜏 = ‫׬‬0 𝑑𝜏 = 2 2
−𝑒 −(𝑡−𝜏) + 2𝑒 −2(𝑡−𝜏) 𝑒 −𝑡 − 𝑒 −2𝑡
1 −𝑡
1 −2𝑡
𝑥 𝑡 = 𝑒 𝐴𝑡 𝑥 0 + 2 − 𝑒 + 2 𝑒
𝑒 −𝑡 − 𝑒 −2𝑡

1 1
𝑥1 𝑡 2𝑒 −𝑡− 𝑒 −2𝑡 𝑒 −𝑡− 2𝑒 −2𝑡 𝑥1 (0) − 𝑒 −𝑡 + 𝑒 −2𝑡
= + 2 2
𝑥2 𝑡 −2𝑒 −𝑡 + 2𝑒 −2𝑡 −𝑒 −𝑡 + 2𝑒 −2𝑡 𝑥2 (0) 𝑒 −𝑡 − 𝑒 −2𝑡

If the initial state is zero, or 𝑥 0 = 0, then 𝑥(𝑡) can be simplified to

1 1
𝑥1 𝑡 − 𝑒 −𝑡 + 𝑒 −2𝑡
= 2 2
𝑥2 𝑡 𝑒 −𝑡 − 𝑒 −2𝑡
STATE SPACE REPRESENTATION Lecture Notes by Dr. Linus. A. Aloo