EEE 2502 CONTROL ENGINEERING III By Linus Aloo MAY 2017

Dedan Kimathi University of Technology, DeKUT.

Bachelor of Science Degree in Electrical and Electronic Engineering, Fifth Year 1st Semester, May-Aug 2017.

EEE2502 CONTROL ENGINEERING III

State variable methods: state space mode representation, similarity transformations, state transition, flow
graphs and canonical forms; geometrical interpretation of Eigen values and Eigen-vectors; time solution by
transition matrix and inverse.

Laplace techniques: canonical forms in relation to controllability and observability; the observer, simple
pole placement design technique.

Discrete systems: sampling, data extrapolators and spectral characteristics; design of Z-transform, impulse
in variance and hold equivalent; block manipulation and pulse transfer functions; Z-transform inversion;
stability and damping; discrete compensator realisation design examples; bilinear transformation; root
space. Frequency domain identification using least squares.

References
1. Modern Control Engineering, Kastuhiko Ogata
2. Discrete Time Control Systems, Kastuhiko Ogata
3. Advanced Control Engineering, Roland S. Burns
4. Modern Control Engineering , P. N. Paraskevopoulos

1
EEE 2502 CONTROL ENGINEERING III By Linus Aloo MAY 2017

EEE2502 CONTROL ENGINEERING III ................................................................................... 1

1.0 STATE VARIABLE METHODS ........................................................................................................... 3
1.1 State space representation ................................................................................................................... 3
1.2 State Equations from Transfer Functions ............................................................................................ 9
1.3 Similarity Transformation ................................................................................................................. 15
1.4 Eigen Values and Eigen Vectors ....................................................................................................... 16
1.5 Repeated Eigen values....................................................................................................................... 18
1.6 Laplace Transform Technique............................................................Error! Bookmark not defined.
1.7 Transfer Function from state space equations ....................................Error! Bookmark not defined.
1.8 Time Solution of state equations ........................................................Error! Bookmark not defined.
1.9 Controllability and Observability .......................................................Error! Bookmark not defined.
1.10 Pole Placement Design Technique ...................................................Error! Bookmark not defined.
1.11 State Observer Design ......................................................................Error! Bookmark not defined.

2
EEE 2502 CONTROL ENGINEERING III By Linus Aloo MAY 2017

1.0 STATE VARIABLE METHODS

1.1 State space representation

State space approach is a generalized time-domain method of modelling, analysing and designing a wide
range of control systems.

The approach can deal with

i. Multiple input, multiple output (MIMO) systems or multivariable systems
ii. Nonlinear and time invariant systems
iii. Alternative controller design approaches

The analysis require

i. Input variables
ii. Output variables
iii. State variables

Definitions

State:
State of dynamic system is the smallest set of variables (state variables) such that knowledge of the variables
at time 𝑡 = 𝑡𝑜 plus knowledge of the inputs for 𝑡 ≥ 𝑡𝑜 , this information determines the behaviour of the
system for any time 𝑡 ≥ 𝑡𝑜 .

State variables:
These are variables making up the smallest set of variables that determine the state of the dynamic system.
Let a dynamic system have 𝑛 variables 𝑥1 , 𝑥2 , … . , 𝑥𝑛 to describe the behaviour of the system while the
input is given for 𝑡 ≥ 𝑡𝑜 and the initial state at 𝑡 = 𝑡𝑜 is specified. If the future state of the system is
determined, then such variables are a set of state variables.

State vector:
The behaviour of a given system is described by 𝑛 state variables and can be considered to be 𝑛 components
of a vector x. Such a vector is called a state vector. A state is thus a vector that determines uniquely the
system state 𝑥(𝑡) for any 𝑡 ≥ 𝑡𝑜 once the state at 𝑡 = 𝑡𝑜 is given and the input for 𝑡 ≥ 𝑡𝑜 is specified.

State space:
The 𝑛 dimensional space whose coordinate axis consists of 𝑥1 , 𝑥2 , … . , 𝑥𝑛 axis is called state space. Any
state can be represented by a point in the state space.

Consider the dynamic system shown in Fig. 1.1

Fig. 1.1MIMO Dynamic System

The above dynamic system has MIMO.

3
EEE 2502 CONTROL ENGINEERING III By Linus Aloo MAY 2017

The state of the system is described by a set of first order differential equations in terms of the state variables
𝑥1 , 𝑥2 , … . , 𝑥𝑛 and input variables 𝑢1 , 𝑢2 , … . , 𝑢𝑟 as

𝑥̇ 1 (𝑡) = 𝑓1 (𝑥1 , 𝑥2 , … . , 𝑥𝑛 ; 𝑢1 , 𝑢2 , … . , 𝑢𝑟 ; 𝑡)
𝑥̇ 2 (𝑡) = 𝑓2 (𝑥1 , 𝑥2 , … . , 𝑥𝑛 ; 𝑢1 , 𝑢2 , … . , 𝑢𝑟 ; 𝑡)
.
(1.1)
.
𝑥̇ 𝑛 (𝑡) = 𝑓𝑛 (𝑥1 , 𝑥2 , … . , 𝑥𝑛 ; 𝑢1 , 𝑢2 , … . , 𝑢𝑟 ; 𝑡)

The output of the system𝑦1 (𝑡), 𝑦2 (𝑡), . . . , 𝑦𝑚 (𝑡) may be a function of input variables, state variables and
time. This may be described as
𝑦1 (𝑡) = 𝑔1 (𝑥1 , 𝑥2 , … . , 𝑥𝑛 ; 𝑢1 , 𝑢2 , … . , 𝑢𝑟 ; 𝑡)
𝑦2 (𝑡) = 𝑔2 (𝑥1 , 𝑥2 , … . , 𝑥𝑛 ; 𝑢1 , 𝑢2 , … . , 𝑢𝑟 ; 𝑡)
.
.
(1.2)

𝑦𝑚 (𝑡) = 𝑔𝑚 (𝑥1 , 𝑥2 , … . , 𝑥𝑛 ; 𝑢1 , 𝑢2 , … . , 𝑢𝑟 ; 𝑡)

Definitions of 𝑢(𝑡), 𝑥(𝑡) and 𝑦(𝑡) would be

𝑥1 (𝑡)
𝑥2 (𝑡)
𝑥(𝑡) = . State vector
.
[𝑥𝑛 (𝑡)]

𝑢1 (𝑡)
𝑢2 (𝑡)
𝑢(𝑡) = . Input vector
.
[𝑢𝑟 (𝑡)]

𝑦1 (𝑡)
𝑦2 (𝑡)
𝑦(𝑡) = . Output vector
.
[𝑦𝑚 (𝑡)]

Using the above vectors, equations (1.1) and (1.2) become

𝑥̇ (𝑡) = 𝑓(𝑥, 𝑢, 𝑡)state equation

(1.3)
𝑦(𝑡) = 𝑔(𝑥, 𝑢, 𝑡)output equation
(1.4)
Time invariant systems
If the vector functions of 𝑓 and 𝑔 do not involve time, then the system is said to be a time invariant
system
Then

4
EEE 2502 CONTROL ENGINEERING III By Linus Aloo MAY 2017

𝑥̇ (𝑡) = 𝑓(𝑥, 𝑢)
(1.5)

𝑦(𝑡) = 𝑔(𝑥, 𝑢)
(1.6)

If equations 1.5 and 1.6 are linear then

𝑥̇ (𝑡) = 𝐴𝑥(𝑡) + 𝐵𝑢(𝑡)

(1.7)

𝑦(𝑡) = 𝐶𝑥(𝑡) + 𝐷𝑢(𝑡)

(1.8)

where A, B, C and D are constant matrices

Time varying systems
If the system is time varying, then equations 1.3 and 1.4 result to

𝑥̇ (𝑡) = 𝐴(𝑡)𝑥(𝑡) + 𝐵(𝑡)𝑢(𝑡)

(1.9)

𝑦(𝑡) = 𝐶(𝑡)𝑥(𝑡) + 𝐷(𝑡)𝑢(𝑡)

(1.10)
where
𝐴(𝑡) – state matrix
𝐵(𝑡) – input matrix
𝐶(𝑡) – output matrix
𝐷(𝑡) – direct transition matrix

Example 1.1
Obtain the state space representation of the system described by

𝑑3 𝑦 𝑑2 𝑦 𝑑𝑦
3
+ 6 2
+ 11 + 6𝑦 = 6𝑢
𝑑𝑡 𝑑𝑡 𝑑𝑡

where 𝑦 is the output and 𝑢 is the input to the system

Solution
Knowledge of 𝑦(0), 𝑦̇ (0) and 𝑦̈ (0) together with the input 𝑢(𝑡) for 𝑡 ≥ 0 determines completely the
future of the system. Thus if 𝑦(𝑡), 𝑦̇ (𝑡) and 𝑦̈ (𝑡) are a set of the state variables then

Defining the state variables as

𝑥1 = 𝑦
𝑥2 = 𝑦̇
𝑥3 = 𝑦̈
Then
𝑥̇ 1 = 𝑦̇ = 𝑥2
𝑥2 = 𝑦̇ = 𝑥3
𝑥3 = 𝑦̈ = −6𝑥1 − 11𝑥2 − 6𝑥3 + 6𝑢

5
EEE 2502 CONTROL ENGINEERING III By Linus Aloo MAY 2017

By use of vector matrix notation, then the 3 first order differential equations can be combined into one as

𝑥̇ 1 0 1 0 𝑥1 0
𝑥̇
[ 2] = [ 0 0 𝑥
1 ] [ 2 ] + [0] 𝑢
𝑥̇ 3 −6 −11 −6 𝑥3 6

The output y is given by

𝑥1
𝑦 = [1 0 0] [𝑥2 ]
𝑥3
Example 1.2

Write the state variable formulation of the parallel RLC network shown in Fig. 1.2
A

𝒗(𝒕)
𝑖(𝑡) = 𝐼𝑠𝑖𝑛 𝜔𝑡

Fig. 1.2. Parallel RLC network

Solution
Applying KCL at node A then
𝑖 = 𝑖𝑅 + 𝑖𝐶 + 𝑖𝐿

𝑣(𝑡) 𝑑𝑣(𝑡) 1
𝐼 sin 𝜔𝑡 = +𝐶 + ∫ 𝑣(𝑡) 𝑑𝑡
𝑅 𝑑𝑡 𝐿

Differentiating with respect to t and rearranging yields

𝑑2 𝑣 1 𝑑𝑣 1 𝜔
2
+ + 𝑣 = 𝐼𝑐𝑜𝑠 𝜔𝑡
𝑑𝑡 𝑅𝐶 𝑑𝑡 𝐿𝐶 𝐶

Choosing
𝑣(𝑡) = 𝑥1 (𝑡)

𝑥̇ 1 (𝑡) = 𝑥2

1 1 𝜔
𝑥̇ 2 (𝑡) = − 𝑥1 − 𝑥2 + 𝐼𝑐𝑜𝑠 𝜔𝑡
𝐿𝐶 𝑅𝐶 𝐶

The vector-matrix differential form of the state equation can be written as

0 1 𝑥 0
𝑥̇ 1 ] [ 1 ] + [𝜔
[ 1 ] = [− 1 𝑥2 𝐼𝑐𝑜𝑠 𝜔𝑡 ]
𝑥̇ 2 −
𝐿𝐶 𝑅𝐶 𝐶
and the output
𝑥1
𝑣 = [1 0] [𝑥 ]
2

6
EEE 2502 CONTROL ENGINEERING III By Linus Aloo MAY 2017

NB
Usually in the circuit problem, the current through the inductor and voltage across the capacitor are
chosen as the state variables.

Example 1.3
Obtain the state space representation for the RLC network shown in Fig. 1.3. Assume that the voltage
across 𝑅2 and the current through 𝑅2 are the output variables denoted as 𝑦1 and 𝑦2 respectively and
therefore obtain the output equations.

L1 L2
i1 i2
R1

C
R2
e1 V
e

Fig. 1.3.
Solution
Let us define:
𝑥1 (𝑡) = 𝑣(𝑡)
𝑥2 (𝑡) = 𝑖1 (𝑡)
𝑥3 (𝑡) = 𝑖2 (𝑡)

The equations governing the behavior of the RLC network are:

𝑑𝑣
𝑖1 + 𝑖2 + 𝐶 =0
𝑑𝑡
𝑑 𝑖1
𝐿1 + 𝑅1 𝑖1 + 𝑒 − 𝑣 = 0
𝑑𝑡
𝑑 𝑖2
𝐿2 + 𝑅2 𝑖2 − 𝑣 = 0
𝑑𝑡
In order to express the variables,
𝑑𝑣 𝑑 𝑖1 𝑑 𝑖2
( ),( ) 𝑎𝑛𝑑 ( )
𝑑𝑡 𝑑𝑡 𝑑𝑡
as linear combinations of the variables 𝑣1, 𝑖1, 𝑖2 & 𝑒 as required in the usual representation, we have:

𝑑𝑣 1 1
= − [ ] 𝑖1 − [ ] 𝑖2
𝑑𝑡 𝐶 𝐶
𝑑 𝑖1 1 𝑅1 1
= ( ) 𝑣 − 𝑖1 − ( ) 𝑒
𝑑𝑡 𝐿1 𝐿1 𝐿1

7
EEE 2502 CONTROL ENGINEERING III By Linus Aloo MAY 2017

𝑑 𝑖2 1 𝑅2
= ( )𝑣 − 𝑖
𝑑𝑡 𝐿2 𝐿2 2

For input, 𝑢(𝑡) = 𝑒(𝑡), we have:

−1 −1
0
𝑥̇ 1 𝐶 𝐶 𝑥1 0
1 −𝑅1 −1
[𝑥̇ 2 ] = 0 [𝑥2 ] + [ ] 𝑢
𝐿1 𝐿1 𝐿1
𝑥̇ 3 1 −𝑅2 𝑥3 6
[𝐿2 0
𝐿2 ]

Assuming that the voltage across 𝑅2 and the current through 𝑅2 as the output variables 𝑦1 and 𝑦2
respectively, the output equations are given by:
𝑥1
𝑦1 0 0 𝑅2 𝑥
⌈𝑦 ⌉ = ⌈ ⌉ ⌈ 2⌉
2 0 0 1 𝑥
3

Exercise
1. Consider the mechanical system shown in Fig.1.4. Assume that the system is a single-input-single-
output linear system, the external force u(t) is the input to the system, and the displacement y(t) of
the mass is the output which is measured from the equilibrium position in the absence of the
external force.
a) Obtain the state space representation of the system.
b) obtain the transfer function for the system from the state-space equations
c) Sketch the block diagram of the system.

.
Fig.1.4. Mechanical System.
2. Consider the system shown in Fig. 1.5. The system involves one integrator and two delayed
integrators. The output of each integrator or delayed integrator can be a state variable. By defining
the output of the plant as 𝑥1 , the output of the controller as 𝑥2 and the output of the sensor as 𝑥3 .
Obtain a state-space model of the system.

8
EEE 2502 CONTROL ENGINEERING III By Linus Aloo MAY 2017

Fig.1.5. Control System.

1.2 State Equations from Transfer Functions

Consider a system defined by

𝑦 (𝑛) + 𝑎1 𝑦 (𝑛−1) + ⋯ + 𝑎𝑛−1 𝑦̇ + 𝑎𝑛 𝑦 = 𝑏0 𝑢𝑛 + 𝑏1 𝑢𝑛−1 + ⋯ + 𝑏𝑛−1 𝑢̇ + 𝑏𝑛 𝑢

(1.11)

where 𝑢 is the input and 𝑦 is the output.

The equation can be written as

𝑌(𝑠)
𝑈(𝑠)
𝑏0 𝑠 𝑛 + 𝑏1 𝑠 𝑛−1 + ⋯ + 𝑏𝑛−1 𝑠 + 𝑏𝑛
= 𝑛 (1.12)
𝑠 + 𝑎1 𝑠 𝑛−1 + ⋯ + 𝑎𝑛−1 𝑠 + 𝑎𝑛

The state space representation of the system defined by equation (1.11) and (1.12) can be presented in
controllable canonical form, observable canonical form and diagonal canonical form

Controllable canonical form

The controllable canonical form of the state space representation is given by

𝑥̇ 1 0 1 0 . . . 0 𝑥1 0
𝑥̇ 2 0 0 1 . . . 0 𝑥 2 0
. . . . . . .
. = . . . . . + . 𝑢 (1.13)
. . . . . . .
𝑥̇ 𝑛−1 0 0 0 . . . 1 𝑥𝑛−1 0
[ 𝑥̇ 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 (1.14)
.
[𝑥𝑛 ]
Example 1.3

9
EEE 2502 CONTROL ENGINEERING III By Linus Aloo MAY 2017

Consider the system given by

𝑌(𝑠) 𝑠+3
= 2
𝑈(𝑠) 𝑠 + 3𝑠 + 2

Obtain the controllable canonical form of the state space representation

Solution

𝑛 = 2 𝑏0 = 0 𝑏1 = 1 𝑏1 = 1 𝑏2 = 3 𝑎1 = 1 𝑎2 = 2

𝑥̇ 0 1 𝑥1 0
[ 1] = [ ][ ] + [ ]𝑢
𝑥̇ 2 −𝑎2 −𝑎1 𝑥2 1

0 1 𝑥1 0
=[ ] [𝑥 ] + [ ] 𝑢
−2 −3 2 1
𝑥
𝑦 = [𝑏2 − 𝑎2 𝑏0 𝑏1 − 𝑎1 𝑏0 ] [𝑥1 ] + 𝑏0 𝑢
2

𝑥1
= [3 1] [𝑥 ]
2

Observable canonical form

For the transfer function

𝑌(𝑠) 𝑏0 𝑠 𝑛 + 𝑏1 𝑠 𝑛−1 + ⋯ + 𝑏𝑛−1 𝑠 + 𝑏𝑛
= 𝑛
𝑈(𝑠) 𝑠 + 𝑎1 𝑠 𝑛−1 + ⋯ + 𝑎𝑛−1 𝑠 + 𝑎𝑛

The observable canonical form of the state space representation is given by

𝑥̇ 1 0 0 . . . 0 −𝑎𝑛 𝑥1
𝑥̇ 2 1 0 . . . 0 −𝑎𝑛−1 𝑥2
. . . . . .
. = . . . . .
. . . . . .
𝑥̇ 𝑛−1 0 0 . . . 0 −𝑎2 𝑥𝑛−1
[ 𝑥̇ 𝑛 ] [0 0 . . . 1 −𝑎1 ] [ 𝑥𝑛 ]
𝑏𝑛 − 𝑎𝑛 𝑏0
𝑏𝑛−1 − 𝑎𝑛−1 𝑏0
.
+ . 𝑢 (1.15)
.
𝑏2 − 𝑎2 𝑏0
[ 𝑏1 − 𝑎1 𝑏0 ]

and

10
EEE 2502 CONTROL ENGINEERING III By Linus Aloo MAY 2017

𝑥1
𝑥2
.
𝑦 = [0 0 . . . 0 1] .
.
𝑥𝑛−1
[ 𝑥𝑛 ]
+ 𝑏0 𝑢 (1.16)
Example 1.4
Obtain the observable canonical form of the state space representation for the system with the transfer
function

𝑌(𝑠) 𝑠+3
= 2
𝑈(𝑠) 𝑠 + 3𝑠 + 2

Solution

𝑛 = 2 𝑏0 = 0 𝑏1 = 1 𝑏1 = 1 𝑏2 = 3 𝑎1 = 1 𝑎2 = 2

𝑥̇ 0 −𝑎2 𝑥1 𝑏 − 𝑎2 𝑏0
[ 1] = [ ][ ] + [ 2 ]𝑢
𝑥̇ 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

Diagonal canonical form

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

= (1.17)
𝑈(𝑠) (𝑠 + 𝑝1 )(𝑠 + 𝑝2 ) . . . (𝑠 + 𝑝𝑛−1 )(𝑠 + 𝑝𝑛 )

1𝑐 2 𝑐 𝑐 𝑛−1 𝑐
𝑛
= 𝑏0 + (𝑠+𝑝 )
+ (𝑠+𝑝 . . . + (𝑠+𝑝 + (𝑠+𝑝 (1.18)
1 2) 𝑛−1 ) 𝑛)

The diagonal canonical form of the state space representation of the system is given by

11
EEE 2502 CONTROL ENGINEERING III By Linus Aloo MAY 2017

𝑥̇ 1 −𝑝1 0 . . . 0 0 𝑥1
𝑥̇ 2 0 −𝑝1 . . . 0 0 𝑥2
. . . . . .
. = . . . . .
. . . . . .
𝑥̇ 𝑛−1 0 0 . . . −𝑝𝑛−1 0 𝑥𝑛−1
[ 𝑥̇ 𝑛 ] [ 0 0 . . . 0 −𝑝𝑛 ] [ 𝑥𝑛 ]
1
1
.
+ . 𝑢 (1.19)
.
1
[1]
and

𝑥1
𝑥2
.
𝑦 = [𝑐1 𝑐2 . . . 𝑐𝑛−1 𝑐𝑛 ] .
.
𝑥𝑛−1
[ 𝑥𝑛 ]
+ 𝑏0 𝑢 (1.20)
Example1.5
Obtain the diagonal canonical form of the state space representation for the system with the transfer
function

𝑌(𝑠) 𝑠+3
= 2
𝑈(𝑠) 𝑠 + 3𝑠 + 2

Solution

𝑠+3 𝑠+3 𝑐1 𝑐2
= = +
𝑠2 + 3𝑠 + 2 (𝑠 + 1)(𝑠 + 2) (𝑠 + 1) (𝑠 + 2)

𝑏0 = 0 𝑐1 = 2 𝑐2 = −1 𝑝1 = 1 𝑝2 = 2

𝑥̇ −𝑝 0 𝑥1 1
[ 1] = [ 1 ][ ] + [ ]𝑢
𝑥̇ 2 0 −𝑝2 𝑥2 1

−1 0 𝑥1 1
=[ ][ ] + [ ]𝑢
0 −2 𝑥2 1
𝑥
𝑦 = [𝑐1 𝑐2 ] [ 1 ] + 𝑏0 𝑢
𝑥2
𝑥1
= [2 −1] [𝑥 ]
2

12
EEE 2502 CONTROL ENGINEERING III By Linus Aloo MAY 2017

Jordan Canonical Form

If the denominator polynomial involves multiple roots then the state space representation can be written in
Jordan canonical form.
For example, if the 𝑝𝑖 ‘s are different from one another except that the first three are equal i.e. 𝑝1 = 𝑝2 =
𝑝3 , then the factored form of the transfer function becomes

𝑌(𝑠)
𝑈(𝑠)
𝑏0 𝑠 𝑛 + 𝑏1 𝑠 𝑛−1 + . . . +𝑏𝑛−1 𝑠 + 𝑏𝑛
= (1.21)
(𝑠 + 𝑝1 )3 (𝑠 + 𝑝4 ) . . . (𝑠 + 𝑝𝑛 )
The partial fraction expansion of the equation becomes

𝑌(𝑠) 𝑐1 𝑐2 𝑐3 𝑐4
= 𝑏0 + 3
+ 2
+ +
𝑈(𝑠) (𝑠 + 𝑝1 ) (𝑠 + 𝑝1 ) (𝑠 + 𝑝1 ) (𝑠 + 𝑝4 )
𝑐𝑛
+. . . + (1.22)
(𝑠 + 𝑝𝑛 )

The state space representation of the system in Jordan canonical form becomes

𝑥̇ 1 −𝑝1 1 0 0 . . . 0 𝑥1 0
𝑥̇ 2 0 −𝑝1 1 . . 𝑥2 0
𝑥̇ 3 0 0 −𝑝1 0 . . . 0 𝑥3 1
𝑥̇ 4 0 −𝑝4 . 𝑥4 1
= . + 𝑢 (1.23)
. . . .
. . . . .
. . . . .
[𝑥̇ 𝑛 ] [ 0 . . 0 . . . −𝑝𝑛 ] [𝑥 𝑛 ] [1 ]

and
𝑥1
𝑥2
.
𝑦 = [𝑐1 𝑐2 . . . 𝑐𝑛 ] . + 𝑏0 𝑢 (1.24)
.
[𝑥𝑛 ]
The characteristic equation of a given system remains invariant under different forms of state variable
representation. This can be true for the transfer function also. The choice of the states is not unique

Example 1.6
Consider the system

𝑦⃛ + 6𝑦̈ + 11𝑦̇ + 6𝑦 = 6𝑢

where 𝑦 is the output and 𝑢 is the input. Obtain the state space representation of the system

Solution I

Let
𝑥1 = 𝑦 𝑥1̇ = 𝑥2
𝑥2 = 𝑦̇ 𝑥̇ 2 = 𝑥3

13
EEE 2502 CONTROL ENGINEERING III By Linus Aloo MAY 2017

𝑥3 = 𝑦̈ 𝑥̇ 3 = −6𝑥1 − 11𝑥2 − 6𝑥3 + 6𝑢

These can be written in the vector matrix differential equation form as

𝑥̇ 1 0 1 0 𝑥1 0
[𝑥̇ 2 ] = [ 0 0 1 ] [𝑥2 ] + [0] 𝑢
𝑥̇ 3 −6 −11 −6 𝑥3 6
𝑥1
𝑦 = [1 0 0] [𝑥2 ]
𝑥3

The characteristic equation becomes

|𝑆𝐼 − 𝐴| = 0

𝑠 3 + 6𝑠 2 + 11𝑠 + 6 = 0

Solution II

𝑦⃛ + 6𝑦̈ + 11𝑦̇ + 6𝑦 = 6𝑢

The transfer function is obtained as

𝑌(𝑠) 6
= 3 2
𝑈(𝑠) 𝑠 + 6𝑠 + 11𝑠 + 6

6
=
(𝑠 + 1)(𝑠 + 2)(𝑠 + 3)
By partial fraction expansion

𝑌(𝑠) 3 −6 3
= + +
𝑈(𝑠) (𝑠 + 1) (𝑠 + 2) (𝑠 + 3)

3 −6 3
𝑌(𝑠) = 𝑈(𝑠) + 𝑈(𝑠) + 𝑈(𝑠)
(𝑠 + 1) (𝑠 + 2) (𝑠 + 3)

Defining 𝑌(𝑠) = 𝑋1 (𝑠) + 𝑋2 (𝑠) + 𝑋3 (𝑠)

where

3
𝑋1 (𝑠) = 𝑈(𝑠)𝑥̇ 1 = −𝑥1 + 3𝑢
(𝑠 + 1)

−6
𝑋2 (𝑠) = 𝑈(𝑠)𝑥̇ 2 = −2𝑥2 − 6𝑢
(𝑠 + 2)

3
𝑋1 (𝑠) = 𝑈(𝑠) 𝑥̇ 3 = −3𝑥3 + 3𝑢
(𝑠 + 3)

14
EEE 2502 CONTROL ENGINEERING III By Linus Aloo MAY 2017

The state space representation becomes

𝑥̇ 1 1 0 0 𝑥1 3
[𝑥̇ 2 ] = [0 −2 0 ] [𝑥2 ] + [−6] 𝑢
𝑥̇ 3 0 0 −3 𝑥3 3
𝑥1
𝑦 = [1 1 1] [𝑥2 ]
𝑥3
The characteristic equation in this case becomes

|𝑆𝐼 − 𝐴| = 0

𝑠 3 + 6𝑠 2 + 11𝑠 + 6 = 0

1.3 Similarity Transformation

It has been stated that the choice of states is not unique for a given system. Suppose that there exists a set
of state variables

𝑋 = [𝑥1 𝑥2 . . . 𝑥𝑛 ]𝑇 (1.25)

We may take another set of state variables

𝑍 = [𝑧1 𝑧 . . . 𝑧𝑛 ]𝑇 (1.26)

So that a linear or similarity transformation exists

Let
𝑋 = 𝑃𝑍 𝑖. 𝑒 𝑍 = 𝑃−1 𝑋 (1.27)

where𝑃 is a non-singular transformational matrix

Differentiating equation (1.27) yields

𝑍̇ = 𝑃−1 𝑋̇ (1.28)

Using the general state space equations

𝑋̇ = 𝐴𝑥 + 𝐵𝑢
𝑦 = 𝐶𝑥

Then equation (1.28) becomes

𝑍̇ = 𝑃−1 𝐴𝑥 + 𝑃−1 𝐵𝑢 (1.29)

From equation (1.27) 𝑋 = 𝑃𝑍

Equation (1.29) becomes

15
EEE 2502 CONTROL ENGINEERING III By Linus Aloo MAY 2017

𝑍̇ = 𝑃−1 𝐴𝑃𝑧 + 𝑃−1 𝐵𝑢 (1.30)

and
𝑦 = 𝐶𝑥
= 𝐶𝑃𝑧 (1.31)

Equations (1.30) and (1.31) can be written as

𝑍̇ = 𝐴̂𝑧 + 𝐵̂ 𝑢 (1.32)

𝑦 = 𝐶̂ 𝑥 z (1.33)
where 𝐴̂ = 𝑃−1 𝐴𝑃, 𝐵̂ = 𝑃−1 𝐵 𝑎𝑛𝑑 𝐶̂ = 𝐶𝑃

Hence similarity transformation the transformed system can be represented in the vector-matrix
differential form as

𝑍̇ = 𝐴̂𝑧 + 𝐵̂𝑢 and the output as

𝑦 = 𝐶̂ 𝑥 z

NB
i. The characteristic equations and hence the Eigen values of 𝐴 and 𝐴̂ are invariant under similarity
transformation
ii. The transfer function remains invariant under similarity transformation

1.4 Eigen Values and Eigen Vectors

Given the matrix equation

𝐴𝑥 = 𝜆𝑥 (1.34)

The values of the scalar 𝜆 for which non trivial solutions exist are called Eigen values and the corresponding
solutions 𝑥 = 0 are called Eigen vectors

Equation (1.34) can be written in the form

(𝜆𝐼 − 𝐴)𝑥 = 0 where I is the identity matrix

|𝜆𝐼 − 𝐴| = 0is the characteristic equation of A

The roots of the characteristic equation are called Eigen values of the matrix A

Corresponding to each Eigen value is a non-zero solution of 𝑥=𝑒𝑖 . This is called the Eigen vector of A
corresponding to 𝜆𝑖

Example 1.7
4 1
Determine the Eigen values and Eigen vectors of 𝐴𝑥 = 𝜆𝑥 where 𝐴 = [ ]
3 2

16
EEE 2502 CONTROL ENGINEERING III By Linus Aloo MAY 2017

Solution
|𝜆𝐼 − 𝐴| = 0

(𝜆 − 4)(𝜆 − 2) − 3 = 0

𝜆1 = 5 , 𝜆2 = 1

For 𝜆1 = 5, 𝐴𝑥 = 𝜆𝑥 becomes

4 1 𝑥1 𝑥1
[ ] [𝑥 ] = 5 [𝑥 ]
3 2 2 2

4𝑥1 + 𝑥2 = 5𝑥1
3𝑥1 + 2𝑥2 = 5𝑥2

𝑥1 = 𝑥2

Eigen vector corresponding to 𝜆1 = 5 is

𝑥1 1
𝑒1 = [𝑥 ] = [ ] simplest form
2 1

For 𝜆2 = 1, 𝐴𝑥 = 𝜆𝑥 becomes

4 1 𝑥1 𝑥1
[ ] [𝑥 ] = 1 [𝑥 ]
3 2 2 2

4𝑥1 + 𝑥2 = 𝑥1
3𝑥1 + 2𝑥2 = 𝑥2

𝑥1 = −3𝑥2

Eigen vector corresponding to 𝜆2 =1is

𝑥1 1
𝑒2 = [𝑥 ] = [ ] simplest form
2 −3

We obtain the transformation matrix 𝑃 from the Eigen vectors as follows

𝑃 = [𝑒1 𝑒2 ] = [1 1
]
1 −3

The matrix 𝐴̂ = 𝑃−1 𝐴𝑃 is then obtained

1 1 −1 4 1 1 1 5 0
𝐴̂ = 𝑃−1 𝐴𝑃 = [ ] [ ][ ]=[ ]
1 −3 3 2 1 −3 0 1

Under similarity transformation the characteristic equation does not change i.e.

|𝜆𝐼 − 𝐴| = |𝜆𝐼 − 𝐴̂| = 𝜆2 + 𝜆𝑠 + 5

17
EEE 2502 CONTROL ENGINEERING III By Linus Aloo MAY 2017

1.5 Repeated Eigen values

In some cases, the matrix 𝐴 will have repeated Eigen values. The Eigen vectors are evaluated as follows

Example 1.8
3 −3 2
Determine the Eigen values and Eigen vectors of 𝐴𝑥 = 𝜆𝑥 where 𝐴 = [−1 5 −2]
−1 3 0

Solution

|𝜆𝐼 − 𝐴| = 0

𝜆−3 3 −2
| 1 𝜆−5 2 |=0
1 −3 𝜆

𝜆3 − 8𝜆2 + 20𝜆 − 16 = 0

𝜆1 = 4, 𝜆2 = 𝜆3 = 2

For 𝜆1 = 4

3 −3 2 𝑥1 𝑥1
[−1 5 −2] [𝑥2 ] = 4 [[𝑥2 ]]
−1 3 0 𝑥3 𝑥3

This yield

𝑥2 =−𝑥1 𝑎𝑛𝑑 𝑥3 = −𝑥1

and the simplest form of the corresponding Eigen vector as

1
𝑒1 = [−1]
−1

For 𝜆2 = 𝜆3 = 2
3 −3 2 𝑥1 𝑥1
[−1 5 −2] [𝑥2 ] = 2 [[𝑥2 ]]
−1 3 0 𝑥3 𝑥3

The simultaneous equations obtained are

𝑥1 − 3𝑥2 + 2𝑥3 = 0

We let 𝑥2 = 𝛼 𝑎𝑛𝑑 𝑥3 = 𝛽 where are 𝛼 and 𝛽 constants

The above equation becomes

𝑥1 − 3𝛼 + 2𝛽 = 0

18
EEE 2502 CONTROL ENGINEERING III By Linus Aloo MAY 2017

𝑥1 = 3𝛼 − 2𝛽and the resulting Eigen vector as

3 −2
𝑒 = 𝛼 [ 1] + 𝛽 [ 0 ]
0 1
3
For 𝛽 = 0 𝑎𝑛𝑑 𝛼 = 1 , 𝑒2 = [1]
0
−2
For 𝛽 = 1 𝑎𝑛𝑑 𝛼 = 0 , 𝑒3 = [ 0 ]
1

The transformation matrix 𝑃 is obtained as

1 3 −2
𝑃 = [−1 1 0 ]
−1 0 1

The matrix 𝐴̂ = 𝑃−1 𝐴𝑃 is then obtained as

1 3 −2 −1 3 −3 2 1 3 −2
𝐴̂ = 𝑃−1 𝐴𝑃 = [−1 1 0 ] [−1 5 −2] [−1 1 0]
−1 0 1 −1 3 0 −1 0 1

4 0 0
= [0 2 0]
0 0 2

Checking the characteristic equation

|𝜆𝐼 − 𝐴| = |𝜆𝐼 − 𝐴̂| = 𝜆3 − 8𝜆2 + 20𝜆 − 16 = 0

19