M.

Renuga / AP – ECE / AAMEC UNIT -5 CONTROL SYSTEM ANALYSIS USING STATE VARIABLE METHODS EC8391 – Control System engg

Syllabus: UNIT V CONTROL SYSTEM ANALYSIS USING STATE VARIABLE METHODS 9

State variable representation-Conversion of state variable models to transfer functions-Conversion of transfer
functions to state variable models-Solution of state equations-Concepts of Controllability and Observability-
Stability of linear systems-Equivalence between transfer function and state variable representations-State
variable analysis of digital control system-Digital control design using state feedback
----------------------------------------------------------------------------------------------------------------------------------
Explain the detail the state representation for continuous time systems (dec–10, 8 marks)
A dynamic system consisting of a finite number of lumped elements may be described by ordinary
differential equations in which time is the independent variable. An n–the order differential equation may be
expressed by a first order vector–matrix differential. In elements of the vector are a set of state variables, then
the vector–matrix differential equation is called a state equation.

Consider the following n–the order system

𝑦 𝑛 + 𝑎1 𝑦 𝑛−1 + … … . . +𝑎𝑛−1 𝑦 + 𝑎𝑛 𝑦 = 𝑢 ….(1)

Noting that the knowledge of y(0), y(0),….., 𝑦0𝑛−1, together with the input u(t) for t≥ 0, determines
completely the future behavior of the system, we may take y(t), y(t),…., 𝑦 𝑛−1 (t) as a set of n state variables.

Let us define

𝑥1 = y ,𝑥2 = 𝑦̇ ………𝑥𝑛 = 𝑦 𝑛−1

Eqn.(1) can be written as 𝑥1̇ = 𝑥2 , 𝑥2̇ = 𝑥3 …… 𝑥̇ 𝑛−1 = 𝑥𝑛

, 𝑥𝑛̇ = −𝑎𝑛 𝑥1 ……−𝑎𝑥 𝑥𝑛−1 + u

Or 𝑥̇ = Ax + Bu …..(2)

Y= Cx –––output equation

The first order differential equation,(2)is the state equation, and the algebraic equation (3) is the
output equation.

Explain in detail the state space representation for discrete time systems. (dec–10, 8 marks)

The discrete form of the state space representation is quite analogous to the continuous form.

The most general state space representation of linear discrete time system is

x (k+1) = G(k) x(k) + H(k) U(k)

y (k) = C(k) x(k) + D(k) U(k)

where x(k) is the state vector, U(k) is input vector and y(k) is the output vector .

The state equation of a discrete time system is a set of n–numbers of first order difference
equations.

𝑥1 (𝑘 + 1) = 𝑓1 (𝑥1 , 𝑥2 , … , 𝑥𝑛 ; 𝑢1 , 𝑢2 ,…., 𝑢𝑚 )

𝑥2 (𝑘 + 1) = 𝑓1 (𝑥1 , 𝑥2 , … , 𝑥𝑛 ; 𝑢1 , 𝑢2 ,…., 𝑢𝑚 )

1
M.Renuga / AP – ECE / AAMEC UNIT -5 CONTROL SYSTEM ANALYSIS USING STATE VARIABLE METHODS EC8391 – Control System engg

𝑥𝑛 (𝑘 + 1) = 𝑓𝑛 (𝑥1 , 𝑥2 , … , 𝑥𝑛 ; 𝑢1 , 𝑢2 ,…., 𝑢𝑚 )

In the matrix form of the above equations can be expressed as

𝑥1 (𝑘 + 1) 𝒂𝟏𝟏 𝒂𝟏𝟐 𝒂𝟏𝒏 𝒙𝟏 𝒃𝟏𝟏 𝒃𝟏𝟐 𝒃𝟏𝒎 𝑼𝟏

(𝑘 𝒂 𝒂𝟐𝟏 𝒂𝟐𝟏 𝒙 𝒃 𝒃𝟐𝟏 𝒃𝟐𝒎 𝑼
[𝑥2 + 1) ] = [ 𝟐𝟏 ] [ 𝟐 ] + [ 𝟐𝟏 ] [ 𝟐]
. . . . . . . . .
𝑥𝑛 (𝑘 + 1) 𝒂𝒏𝟏 𝒂𝒏𝟐 𝒂𝒏𝒏 𝒙𝒏 𝒃𝒏𝟏 𝒃𝒏𝟐 𝒃𝒏𝒎 𝑼𝒎

The above matrix equation can be written in the vector notation as

𝑥1 (𝑘 + 1) = 𝐴 𝑥(𝑘) + 𝐵 𝑈(𝑘) …(1)

𝑤ℎ𝑒𝑟𝑒 𝐴 = 𝑠𝑦𝑠𝑡𝑒𝑚 𝑚𝑎𝑡𝑟𝑖𝑥 𝑜𝑓 𝑜𝑟𝑑𝑒𝑟(𝑛 × 𝑛)

𝐵 = 𝑖𝑛𝑝𝑢𝑡 𝑚𝑎𝑡𝑟𝑖𝑥 𝑜𝑓 𝑜𝑟𝑑𝑒𝑟(𝑛 × 𝑚)

The Eqn.(1) is the state equation of discrete system.

The output at any discrete time instant, k are functions of state variables and inputs. Hence the
output variables of linear time invariant system can be expressed as a linear combination of state
variables and inputs.

In the matrix form,

𝑦1 𝒄𝟏𝟏 𝒄𝟏𝟐 𝒄𝟏𝒏 𝒙𝟏 𝒅𝟏𝟏 𝒅𝟏𝟐 𝒅𝟏𝒎 𝑼𝟏

𝑦 𝒄 𝒄𝟐𝟏 𝒄𝟐𝟏 𝒙 𝒅 𝒅𝟐𝟏 𝒅𝟐𝒎 𝑼
[ 2 ] = [ 𝟐𝟏 ] [ 𝟐 ] + [ 𝟐𝟏 ] [ 𝟐]
. . . . . . . . .
𝑦𝑛 𝒄𝒏𝟏 𝒄𝒏𝟐 𝒄𝒏𝒏 𝒙𝒏 𝒅𝒏𝟏 𝒅𝒏𝟐 𝒅𝒏𝒎 𝑼𝒎

The above matrix equation can be written in vector notation as

𝑦(𝑘) = 𝐶 𝑥(𝑘) + 𝐷 𝑈(𝑘) …(2)

Where, C = output matrix of order (p × n)

D = output matrix of order (p × m)

𝒅𝟑 𝒚 𝒅𝟐 𝒚 𝒅𝒚
1. Construct a state model for a system characterized by the differential equation, 𝒅𝒕𝟑
+ 6 𝒅𝒕𝟐
+11 𝒅𝒕 +6y+u=0.
Give the block diagram representation of the state model.
Solution
Let us choose y and their derivatives as state variables. The system is governed by third order
differential equation and so the number of state variable are three.
The state variables 𝑥1 , 𝑥2 and 𝑥3 are related to phase variables as follows
𝑥1 = y
𝑑𝑦
𝑥2 = = 𝑥1
𝑑𝑡
𝑑2 𝑦
𝑥3 = = 𝑥2
𝑑𝑡 2

𝑑𝑦 𝑑2 𝑦 𝑑3 𝑦
Put y = 𝑥1 , = 𝑥2 and = 𝑥3 and = 𝑥̇ 3 in the given equation,
𝑑𝑡 𝑑𝑡 2 𝑑𝑡 3
∴ 𝑥̇ 3 + 6𝑥3 + 11𝑥2 + 6𝑥1 + u =0 (or) 𝑥̇ 3 = – 6𝑥3 – 11𝑥2 – 6𝑥1 – u

The state equation are

𝑥̇ 1 = 𝑥2
𝑥̇ 2 = 𝑥3
𝑥̇ 3 = – 6𝑥1 – 11𝑥2 – 6𝑥3 – u
On arranging the state equations in the matrix form we get,
𝑥̇ 1 0 1 0 𝑥1 0
[ 𝑥̇ 2 ] = [ 0 0 1] 𝑥
[ 2 ] + [ 0 ] [−𝑢]
𝑥̇ 3 −6 −11 −6 𝑥3 −1

Here, y = output ; But, y = 𝑥1

2
M.Renuga / AP – ECE / AAMEC UNIT -5 CONTROL SYSTEM ANALYSIS USING STATE VARIABLE METHODS EC8391 – Control System engg

𝑥1
∴ The output equation is, y = [1 0 0] [ 𝑥2 ]
𝑥3
The state equation and output equation, constitutes the state model of the system.

The block diagram from of the state diagram of the system is shown in fig

u -u 𝑥̇ 3 x3 = 𝑥̇ 2 x2 = 𝑥̇ 1 y = x1
-1 ∫ ∫ ∫

6x3
6

11x2
11

6x1
6
𝒀(𝒔) 𝟏𝟎
2. Obtain the state model of the system whose transfer function is given as, =
𝑼(𝒔) 𝒔𝟑 +𝟒𝒔𝟐 +𝟐𝒔+𝟏

Solution:
𝑌(𝑠) 10
Given that = ......(1)
𝑈(𝑠) 𝑠3 +4𝑠2 +2𝑠+1
On cross multiplying the Eqn.(1) we get,
Y(s)[ s3 + 4s 2 + 2s + 1] = 10 U(s)
Y(s) s3 Y(s) + 4s 2 Y(s) + 2s Y(s) + Y(s) = 10 U(s) ......(2)
On taking inverse laplace transform of Eqn.(2) we get,
𝑦⃛ + 4𝑦̈ + 2𝑦̇ + y = 10u .......(3)

Let us define state variable as follows,

𝑥1 = 1 ; 𝑥2 = 𝑦̇ ; 𝑥3 = 𝑦̈
Put 𝑦⃛ = 𝑥3̇ ; 𝑦̈ = 𝑥3 ; 𝑦̇ = 𝑥2 and y = 𝑥1 in the Eqn.(3)
∴ 𝑥3̇ + 4𝑥3 + 2𝑥2 + 𝑥1 = 10u (or) 𝑥3̇ = –4𝑥3 – 2𝑥2 – 𝑥1 + 10u

The state equation are

𝑥1̇ = 𝑥2 ; 𝑥2̇ = 𝑥3 ; 𝑥3̇ = –4𝑥3 –2𝑥2 – 𝑥1 +10u
The output equation is y = 𝑥1
The state model in the matrix form is,

𝑥̇ 1 0 1 0 𝑥1 0
[ 𝑥̇ 2 ] = [0 0 𝑥
1 ] [ 2 ]+[ 0 ] [𝑢]
𝑥̇ 3 −1 −2 −4 𝑥3 10
𝑥1
Y = [1 0 0] [ 2 ] 𝑥
𝑥3
-----------------------------------------------------------------------
3. Determine the canonical state model of the system, whose transfer function is
T(s)=2(s+5)/[(s+2)(s+3)(s+4)]

Solution
By partial fraction expansion,

Y(s) 2(s+5) A B C
= (s+2)(s+3)(s+4)
= (s+2)
+ (s+3)
+ (s+4)
U(s)

2(s+5) 2(−2+5) 2×3

A = (s+3) (s+4)
|s = –2 = (−2+3) (−2+4)
= = 3
1×2
2(s+5) 2(−3+5) 2×2
B = (s+2) (s+4)
|s = –3 = (−3+2) (−3+4)
= = –4
−1×1

3
M.Renuga / AP – ECE / AAMEC UNIT -5 CONTROL SYSTEM ANALYSIS USING STATE VARIABLE METHODS EC8391 – Control System engg

2(s+5) 2(−4+5) 2×1

C = (s+2) (s+3)
|s = –4 = (−4+2) (−4+3)
= = 1
−2×(−1)
Y(s) 3 4 1
= – + ........(1)
U(s) (s+2) (s+3) (s+4)

The equation 1 can be rearranged as shown below

Y(s) 3 4 1
= – +
U(s) s(1+2⁄𝑠) s(1+3⁄𝑠) s(1+4⁄𝑠)

1 1 1

Y(s) = [ 1
𝑠
x 3]U(s) – [ 𝑠
1 x 4]U(s) + [ 𝑠
1 ] U(s) .......(2)
1+ × 2 1+ ×3 1+ ×4
𝑠 𝑠 𝑠

The Eqn.(2) can be represented by the block diagram in Fig.

Assign state variable at the output of the integrators as shown in fig. At the input of the integrators we
have first derivative of the state variables.
The state equations are formed by adding all the incoming signals to the integrator and equating to the
corresponding first derivative of state variable.

The state equations are: 𝑥1̇ = –2𝑥1 +u ; 𝑥2̇ = –3𝑥2 +u ; 𝑥3̇ = –4𝑥3 +u
𝑥1̇ 𝑥1
1⁄ 3
U(S) 𝑠 Y(s)

1⁄ 3
𝑠 4

𝑥3̇ 3 𝑥3

1⁄
𝑠

The output equation is, y = 3𝑥1 – 4𝑥2 + 𝑥3

The state model in matrix form is given below

𝑥1̇ −2 0 0 𝑥1 1
[𝑥2̇ ] = [ 0 −3 0 ] [ 𝑥2 ]+[ 1] [𝑢]
𝑥3̇ 0 0 −4 𝑥3 1
𝑥1
Y = [3 −4 1] [ 𝑥2 ]
𝑥3
𝟎 𝟏
4. Consider the matric A = [ ]. Compute 𝒆𝑨𝒕 .
−𝟐 −𝟑
Solution
0 1
A = [ ]
−2 −3
eAt = 𝜑(t) = 𝐿−1 [(sI–A)−1]

1 0 0 1 𝑠 0 0 1 𝑠 −1
sI–A = s[ ]–[ ] = [ ] – [ ] = [ ]
0 1 −2 −3 0 𝑠 −2 −3 2 𝑠+3
Let, ∆ = |sI–A| = determinant of (sI–A)

𝑠 −1
∴ ∆ = |sI–A| = | | = s(s+3) + 2 = 𝑠 2 +3s+2 = (s+2)(s+1)
2 𝑠+3
[Confactor of (sI−A)]𝑇 [Confactor of (sI−A)]𝑇
𝜑(s) = [sI–a]−1 = =
𝑑𝑒𝑡𝑒𝑟𝑚𝑖𝑛𝑎𝑛𝑡 𝑜𝑓 (𝑠𝐼−𝐴) ∆

4
M.Renuga / AP – ECE / AAMEC UNIT -5 CONTROL SYSTEM ANALYSIS USING STATE VARIABLE METHODS EC8391 – Control System engg

1 𝑠+3 1
𝜑(𝑠) = [ ]
(𝑠+1)(𝑠+2) −2 𝑠
𝑠+3 1
(𝑠+1)(𝑠+2) (𝑠+1)(𝑠+2)
𝜑(𝑠) = [ −2 𝑠 ]
(𝑠+1)(𝑠+2) (𝑠+1)(𝑠+2)

𝐴1 𝑩1 𝐴2 𝐵2
+ +
(𝑠+1) (𝑠+2) (𝑠+1) (𝑠+2)
𝜑(𝑠) = [ 𝐴3 𝐵3 𝐴4 𝐵4 ]
+ +
(𝑠+1) (𝑠+2) (𝑠+1) (𝑠+2)

s+3 𝐴1 𝐵1
= +
(s+1)(s+2) 𝑠+1 𝑠+2

𝑠+3
𝐴1 = |s=–1 = 2
𝑠+2

𝑠+3
𝐵1 = |s=–2 = –1
𝑠+1

1 𝐴2 𝐵2
= +
(s+1)(s+2) 𝑠+1 𝑠+2

1
𝐴2 = |s=–1 = 1
𝑠+2

1
𝐵2 = |s=–2 = –1
𝑠+1

−2 𝐴3 𝐵3
= +
(s+1)(s+2) 𝑠+1 𝑠+2

−2
𝐴3 = |s=–1 = –2
𝑠+2

−2
𝐵3 = |s=–2 = 2
𝑠+1

𝑠 𝐴4 𝐵4
= +
(s+1)(s+2) 𝑠+1 𝑠+2

𝑠
𝐴4 = |s= –1 = –1
𝑠+2

𝑠
𝐵4 = |s= –2 = 2
𝑠+1

2 1 1 1
− −
∴ 𝜑(𝑠) = [𝑠+1
−2
𝑠+2
2
𝑠+1
−1
𝑠+2
2 ]
+ +
𝑠+1 𝑠+2 𝑠+1 𝑠+2

On taking inverse Laplace transform of 𝜑(𝑠) we get 𝜑(𝑡), where 𝜑(𝑡)=𝑒 𝐴𝑡

−𝑡 −2𝑡
𝑒 𝐴𝑡 = 𝜑(𝑡) = [ 2𝑒 −𝑡 − 𝑒 −2𝑡 𝑒 −𝑡 − 𝑒 −2𝑡 ]
−2𝑒 + 2𝑒 −𝑒 −𝑡 + 2𝑒 −2𝑡

5. For a system represented by state equation 𝐗̇(t) = AX(t). The response is X(t) = [
𝒆−𝟐𝒕 ] when X(0) = [ 𝟏 ]
−𝟐𝒆−𝟐𝒕 −𝟐
−𝒕 𝟏
and X(t) =[ 𝒆 −𝒕 ] when X(0) = [ ]. Determine the system matrix A and the state transition matrix.(Nov/Dec –10)
−𝒆 −𝟏
Solution
The solution of state equation is, X(t) = eAt X(0) ...........(1)
Premultiply the Eqn.(1) by e−At
e−At X(t) = e−At eAt X(0)

∴ 𝑒 −𝐴𝑡 X(t) = X(0) ...........(2)

−2t 1
One of the response is X(t) = [ e −2t ] and X(0) = [ ]
−2e −2
On substituting the response in Eqn.(2) we get,
−2t 1
e−At [ e −2t ] = [ ] ......... (3)
−2e −2
−At
𝑒11 𝑒12
Let, e = [𝑒 ] ......... (4)
21 𝑒22

5
M.Renuga / AP – ECE / AAMEC UNIT -5 CONTROL SYSTEM ANALYSIS USING STATE VARIABLE METHODS EC8391 – Control System engg

From Eqn.(3) and (4) we can write

𝑒11 𝑒12 e−2t ] = [ 1 ]
[𝑒 ] [ ......... (5)
21 𝑒22 −2e−2t −2

On multiplying the Eqn.(5): we get the following two equations,

𝑒11 e−2t − 2𝑒12 e−2t = 1 ......... (6)
𝑒21 e−2t − 2𝑒22 e−2t = –2 ......... (7)
−t 1
The second solution of state equation is X(t) = [ e −t ] and X(0) = [ ]
−e −1
On substituting this solution in equation 2 we get,

−t 1
e−At [ e −t ] = [ ] ......... (8)
−e −1
From Eqn.(4) and (8) we can write,
𝑒11 𝑒12 e−t ] = [ 1 ]
[𝑒 ] [ ......... (9)
21 𝑒22 −e−t −1
On multiplying the Eqn.(9) we get the following two equation,
𝑒11 e−t − 𝑒12 e−t = 1 ......... (10)
𝑒21 e−t − 𝑒22 e−t = –1 ......... (11)

Eqn. (10) × 𝑒 −𝑡 ⇒ 𝑒11 e−2t − 𝑒12 e−2t = 𝑒 −𝑡

Eqn. (6) × 1 ⇒ 𝑒11 e−2t − 2𝑒12 e−2t = 1
(–) (+) (–)
−2𝑡
On subtracting 𝑒12 𝑒 = 𝑒 −𝑡 –1 ....... (12)
From Eqn.(12) we get,

𝑒 −𝑡 −1 𝑒 −𝑡 1
𝑒12 = = – 𝑒 −2𝑡 = 𝑒 𝑡 – 𝑒 2𝑡 ....... (13)
𝑒 −2𝑡 𝑒 −2𝑡

1+2𝑒12 𝑒 −2𝑡
From Eqn. (6), 𝑒11 = 𝑒 −2𝑡

1+2( 𝑒 𝑡 −𝑒 2𝑡 )𝑒 −2𝑡 1+2 𝑒 −𝑡 −2 2 𝑒 −𝑡 −1 2 𝑒 −𝑡 1

Put 𝑒12 = 𝑒 𝑡 – 𝑒 2𝑡 , ∴ 𝑒11 = = = = – = 2𝑒 𝑡 – 𝑒 2𝑡
𝑒 −2𝑡 𝑒 −2𝑡 𝑒 −2𝑡 𝑒 −2𝑡 𝑒 −2𝑡

Eqn. (11) × 𝑒 −𝑡 ⇒ 𝑒21 e−2t − 𝑒22 e−2t = –𝑒 −𝑡

Eqn. (7) × 1 ⇒ 𝑒21 e−2t − 2𝑒22 e−2t = –2

(–) (+) (+)
−2t
On subtracting e22 e = 2–e−t ....... (14)

From Eqn.(14) we get,

−e−t +2 −e−t 2
e22 = = + = –et +2e2t
e−2t e−2t e−2t

−1+e22 e−t
From Eqn.(11): e21 = e−t

−1+(−et +2e2t )e−t

Put e22 = –et +2e2t , ∴ e21 = e−t
−1−1+2et −2 2et
e21 = = + e−t = –2et +2e2t
e−t e−t

𝑒11 𝑒12 2𝑒 𝑡 − 𝑒 2𝑡 𝑒 𝑡 − 𝑒 2𝑡 ]
∴ e−At = [𝑒 𝑒22 ] = [
21 −2𝑒 𝑡 + 2𝑒 2𝑡 −𝑒 𝑡 + 2𝑒 2𝑡
−𝑡 −2𝑡
∴ eAt = [ 2𝑒 −𝑡 − 𝑒 −2𝑡 𝑒 −𝑡 − 𝑒 −2𝑡 ] ; where, eAt is the state transition matrix.
−2𝑒 + 2𝑒 −𝑒 −𝑡 + 2𝑒 −2𝑡

We know that, L[eAt ] = 𝜑(s)

Where 𝜑(s) = (sI–A)−1 ; ∴𝜑(s)−1 = (sI–A) or A = sI – 𝜑(s)−1

2 1 1 1
− −
𝜑(s) = L[e ] = At
[𝑠+1
−2
𝑠+2
2
𝑠+1
−1
𝑠+2
2 ]
+ +
𝑠+1 𝑠+2 𝑠+1 𝑠+2

6
M.Renuga / AP – ECE / AAMEC UNIT -5 CONTROL SYSTEM ANALYSIS USING STATE VARIABLE METHODS EC8391 – Control System engg

2(𝑠+2)−(𝑠+1) (𝑠+2)−(𝑠+1) 𝑠+3 1

(𝑠+1)(𝑠+2) (𝑠+1)(𝑠+2) (𝑠+1)(𝑠+2) (𝑠+1)(𝑠+2)
= [−2(𝑠+2)+2(𝑠+1) −(𝑠+2)+2(𝑠+1)
] = [ −2 𝑠 ]
(𝑠+1)(𝑠+2) (𝑠+1)(𝑠+2) (𝑠+1)(𝑠+2) (𝑠+1)(𝑠+2)

𝑠(𝑠+3)+2 𝑠2 +3𝑠+2 (𝑠+1)(𝑠+2) 1

Determinant of 𝜑(𝑠) = (𝑠+1)2 (𝑠+2)2 = = =
(𝑠+1)2 (𝑠+2)2 (𝑠+1)2 (𝑠+2)2 (𝑠+1)(𝑠+2)

𝑠 −1
(𝑠+1)(𝑠+2) (𝑠+1)(𝑠+2) 𝑠 −1
𝜑(s)−1 = (s+1)(s+2) [ 2 𝑠+3 ] =[ ]
2 𝑠+3
(𝑠+1)(𝑠+2) (𝑠+1)(𝑠+2)

1 0 𝑠 −1 𝑠 0 𝑠 −1 0 1
A = sI – 𝜑(s)−1 =s[ ]–[ ] =[ ] –[ ]=[ ]
0 1 2 𝑠+3 0 𝑠 2 𝑠+3 −2 −3
Result
𝑡 −2𝑡
A=[
0 1
] ; eAt = [ 2𝑒−𝑡− 𝑒 −2𝑡 𝑒 −𝑡 − 𝑒 −2𝑡 ]
−2 −3 −2𝑒 + 2𝑒 −𝑒 −𝑡 + 2𝑒 −2𝑡
------------------------------------------------------------------------
Transformation of state model
1. A linear time invariant system is described by the following state model.
𝑥1 0 1 0 𝑥1 0 𝑥1
𝑥
[ 2] = [ 0 0 1] 𝑥
[ 2 ] + [0 ] u and y = [1 0 0] 𝑥
[ 2]
𝑥3 −6 −11 −6 𝑥3 2 𝑥3
Transform this state model into a canonical state model. Also compute the state transition matrix, 𝑒 𝐴𝑡 .
Solution
To find eigenvalues:
1 0 0 0 1 0 𝜆 0 0 0 1 0 𝜆 −1 0
[𝜆𝐼 − 𝐴] = λ [0 1 0] − [ 0 0 1 ] = [0 𝜆 0 ] − [ 0 0 1 ] = [0 𝜆 −1 ]
0 0 1 −6 −11 −6 0 0 𝜆 −6 −11 −6 6 11 𝜆 + 6
The characteristic equation is, |𝜆I – A| = 0
𝜆 −1 0
(𝜆𝐼 − 𝐴) = [0 𝜆 −1 ] = λ[λ(λ+6)+11]+1 × 6 = 0
6 11 𝜆 + 6
λ (λ2+6λ+11) +6 = 0 ; λ3 + 6λ2 +11λ+ 6 = 0
λ = –1 is one of the root of the above equation,
∴(λ+1)( λ2+5λ+6) = 0
(λ+1)( λ+2)( λ+3) = 0
(λ+1) = 0 ; ( λ+2) = 0 ; ( λ+3) = 0
The eigenvalues are: λ1 = –1 ; λ2 = –2 ; λ3 = – 3
To find modal matrix:

The system is given in the comparision form and therefore modal matrix, M is given by vander monde
matrix, v.
1 1 1 1 1 1
M=V =[ 1 𝜆 𝜆 2 𝜆 3 ] = [−1 −2 −3]
𝜆12 𝜆22 𝜆23 1 4 9
Alternatively model matrix can be found from eigen vectors. The alternate method is shown below,
1 0 0 0 1 0 −1 −1 0
(λ1I – A) = (–1) [0 1 0] − [ 0 0 1 ] = [ 0 −1 −1]
0 0 1 −6 −11 −6 6 11 5
Let the co factors of ( λII – A)alons 1 row be c11,c12, and c13
st

−1 −1 0 −1 0 −1
C11=(+1)| | = 6; C12=(–1)| | = −6; c13=(+1)| |=6
11 5 6 5 6 11
𝐶11 6 1
Eigen value corresponding to 𝜆1 = m11 = [𝐶12 ] = [ −6 ] = [−1]
𝐶13 6 1
Note: the elements of eigen vector can be divided by a constant.
7
M.Renuga / AP – ECE / AAMEC UNIT -5 CONTROL SYSTEM ANALYSIS USING STATE VARIABLE METHODS EC8391 – Control System engg

1 0 0 0 1 0 −2 −1 0
(λ2I – A) = (–2) [0 1 0] − [ 0 0 1 ] = [ 0 −2 −1]
0 0 1 −6 −11 −6 6 11 4
Let the cofactors (λ2I – A) along 1st row be c11,c12, and c13
−2 −1 0 −1 0 −2
C11=(+1)| | = 3; C12=(–1)| | = −6; c13=(+1)| | = 12
11 4 6 4 6 11
𝐶11 3 1
Eigen vector corresponding toλ2 = m2 = [𝐶12 ] =[−6] =[−2]
𝐶13 12 4
1 0 0 0 1 0 −3 −1 0
(λ3I – A) = (–3) [0 1 0] − [ 0 0 1 ] = [ 0 −3 −1]
0 0 1 −6 −11 −6 6 11 3
Let the cofactors (λ3I – A) along 1st row be c11,c12, and c13
−3 −1 0 −1 0 −3
C11=(+1)| | = 2; C12=(–1)| | = −6; c13=(+1)| | = 18
11 3 6 3 6 11
𝐶11 2 1
Eigen vector corresponding toλ3 = m3 = [𝐶12 ] = [−6] = [−3]
𝐶13 18 9
1 1 1
The modal matrix, M=[𝑚1 𝑚2 𝑚3] = [−1 −2 −3]
1 4 9
It is observed that the modal matrix obtained from vander monde matrix and from eigen vectors are same.
To find M–1
(𝐂𝐨𝐟𝐚𝐜𝐭𝐨𝐫 𝐨𝐟 𝐌)𝐓 𝐌cof𝐓
M–1 = 𝐃𝐞𝐭𝐞𝐫𝐦𝐢𝐧𝐚𝐧𝐭 𝐨𝐟 𝐌 = ∆𝐦
1 1 1
∆M = |−1 −2 −3| = 1(–18+12)–1(–9+3)+1(–4+2)=–2
1 4 9
−6 6 −2 −6 −5 −1
𝑴𝑻𝒄𝒐𝒇 = transpose of [−5 8 −3] = [ 6 8 2]
−1 2 −1 −2 −3 −1
−6 −5 −1
–1 1
∵ 𝑴 = [6 8 2]
−2
−2 −3 −1
To find canonical form of state model

−6 −5 −1 0 1 0 1 1 1
–1 1
Grammian matrix( 𝚲 =M AM) = −2 [ 6 8 2][0 0 1 ] [−1 −2 −3]
−2 −3 −1 −6 −11 −6 1 4 9
6 5 1 1 1 1 −1 0 0
1
= −2 [−12 −16 −4] [−1 −2 −3] = [ 0 −2 0 ]
6 9 3 1 4 9 0 0 −3
−6 −5 −1 0 1
1
̃ =M–1 B
𝐁 = −2 [ 6 8 2 ] 0 = −2
−2 −3 −1 2 1
1 1 1
𝐂̃ = CM = [1 0 0] [−1 −2 −3] = [1 1 1]
1 4 9
The canonical form of state model is,
Ż = 𝚲Z+𝐁̃U

Y = 𝐂̃ Z+DU (here DU is not defined)

ż 1 −1 0 0 z1 1 z1
ż 2 = [ 0 −2 0 ] [z2] + [−2] [U] ; Y = [1 1 1] [z2]
ż 3 0 0 −3 z3 1 z3

To compute state transition matrix, 𝒆𝑨𝒕

The state transition matrix, 𝑒 𝐴𝑡 = 𝑀𝑒 Λ𝑡 M–1

8
M.Renuga / AP – ECE / AAMEC UNIT -5 CONTROL SYSTEM ANALYSIS USING STATE VARIABLE METHODS EC8391 – Control System engg

𝑒 𝜆1𝑡 0 0 𝑒 −𝑡 0 0
𝑒 Λ𝑡 = [ 0 𝑒 𝜆2𝑡 0 ]= [ 0 𝑒 −2𝑡 0 ]
0 0 𝑒 𝜆3𝑡 0 0 𝑒 −3𝑡
1 1 1 𝑒 −𝑡 0 0 −6 −5 −1
𝐴𝑡 1
∴𝑒 =[−1 −2 −3] [ 0 𝑒 −2𝑡
0 ] (−2) [ 6 8 2]
1 4 9 −3𝑡 −2 −3 −1
0 0 𝑒

𝑒 −𝑡 𝑒 −2𝑡 𝑒 −3𝑡 3 2.5 0.5

∴ 𝑒 𝐴𝑡 =[−𝑒 −𝑡 −2𝑒 −2𝑡 −3𝑒 −3𝑡 ] [−3 −4 −1]

𝑒 −𝑡 4𝑒 −2𝑡 9𝑒 −3𝑡 1 1.5 0.5

3𝑒 −𝑡 − 3𝑒 −2𝑡 + 𝑒 −3𝑡 2.5𝑒 −𝑡 −4𝑒 −2𝑡 +1.5𝑒 −3𝑡 0.5𝑒 −𝑡 − 𝑒 −2𝑡 +0.5𝑒 −3𝑡
= [−3𝑒 −𝑡 +6𝑒 −2𝑡 −3𝑒 −3𝑡 −2.5𝑒 −𝑡 +8𝑒 −2𝑡 −4.5𝑒 −3𝑡 −0.5𝑒 −𝑡 +2𝑒 −2𝑡 −1.5𝑒 −3𝑡 ]
3𝑒 −𝑡 −12𝑒 −2𝑡 +9𝑒 −3𝑡 2.5𝑒 −2𝑡 −16𝑒 −2𝑡 13.5𝑒 −3𝑡 0.5𝑒 −𝑡 −4𝑒 −2𝑡 +4.5𝑒 −3𝑡
2.Convert the following system matrix to canonical form and hence calculate the state transition matrix
𝒆𝑨𝒕 .
4 1 −2
A = [1 0 2]
1 −1 3
SOLUTION
To find eigenvalues
1 0 0 4 1 −2 𝜆 − 4 −1 2
(𝜆𝐼 − 𝐴) = λ [0 1 0] − [1 0 2 ] = [ −1 𝜆 −2 ]
0 0 1 1 −1 3 −1 1 𝜆−3
𝜆 − 4 −1 2
(𝜆𝐼 − 𝐴) = [ −1 𝜆 −2 ] = (𝜆 − 4)[λ(λ–3)+2]+1(–λ+3–2)+2(–1+λ)
−1 1 𝜆−3
= (λ–4)(λ2–3λ+2)–λ+1–2+2λ=(λ–4)(λ–1)(λ–2)+(λ–1)

= (λ–1)[(λ–4)(λ–2)+1]=(λ–1)[λ2=6λ+8+1]

= (λ–1)(λ2–6λ+9)=(λ–1)(λ–3)2

The eigen values are λ1 =1, λ2=3, λ3=3. Since one of the eigen value has a multiciplicity of 2(repeats two
times) the canonical form will be a Jordan canonical form.

To find eigen vectors

1 0 0 4 1 −2 −3 −1 2
[λ1I–A]=(1) [0 1 0] − [1 0 2 ] = [−1 1 −2]
0 0 1 1 −1 3 −1 1 −2
Let the co factors of ( λ1I – A)alons 1st row be c11,c12, and c13.

1 −2 −1 −2 −1 −1
C11= (+1)| | = 0; C12= (–1)| | = 0; c13= (+1)| |=0
1 −2 −1 −2 −1 1
𝐶11 0
∴ m1=[𝐶12 ]= [ 0 ] the cofactors along 1st row gives null solution.
𝐶13 0

Let the co factors of ( λ1I – A) along the IInd row be c21,c22, and c23

−1 2 3 2 −3 −1
C21=(–1)| | = 0; C12=(+1)| | = 8; c13=(–1)| |=4
1 −2 −1 −2 −1 1
𝐶21 0
∴ m1=[𝐶22 ]= [ 8 ]
𝐶23 4
1 0 0 4 1 −2 𝜆−4 −1 2
[λ2I – A]= (λ2) [0 1 0] − [1 0 2 ] = [ −1 𝜆 −2 ]
0 0 1 1 −1 3 −1 1 𝜆−3
Let the cofactors (λ2I – A) along 1st row be c11,c12, and c13

9
M.Renuga / AP – ECE / AAMEC UNIT -5 CONTROL SYSTEM ANALYSIS USING STATE VARIABLE METHODS EC8391 – Control System engg

𝜆 −2
C11=(+1)| 2 | =λ2(λ2–3)+2=λ22–3λ2+2 ;
1 𝜆2 − 3

−1 −2
C12=(–1)| | = (–1)[(–1)(λ2–3)–2]=λ2–1
−1 𝜆2 − 3

−1 𝜆2
c13=(+1)| | = –1+λ2=λ2–1
−1 1
Let m2 be the independent eigen vector corresponding to λ2=3.

𝐶11 𝜆22 − 3𝜆2 + 2 32 − (3 ∗ 3) + 2 2

Now, m2 = [𝐶12 ] =[ 𝜆2 − 1 ] =[ 3−1 ]= [2]
𝐶13 𝜆2 − 1 3−1 2

The eigenvalue m3 is given by,

𝑑 𝑑
𝐶11 (𝜆22 − 3𝜆2 + 2)
𝑑𝜆2 𝑑𝜆2
𝑑 𝑑
2𝜆2 − 3 2∗3−3 3
m3 = 𝐶12 = (𝜆2 − 1) =[ 1 ]= [ 1 ]= [1]
𝑑𝜆2 𝑑𝜆2
𝑑 𝑑 1 1 1
[ 𝑑𝜆3 𝐶13 ] [ 𝑑𝜆3
(𝜆2 − 1) ]

To find canonical form of system matrix

0 2 3
The modal matrix is given by, M=[𝑚1 𝑚2 𝑚 3 ] = [8 2 1]
4 2 1
To find M–1

0 2 3
–1 𝐌cof𝐓
M = ; ∆M = |8 2 1| = –8+24=16
∆𝒎
4 2 1

0 −4 8 T 0 4 −4
𝑴𝑻𝒄𝒐𝒇 = [ 4 −12 8 ] = [−4 −12 24 ]
−4 24 −16 8 8 −16
0 4 −4 0 1 −1
1 1
∴ 𝑴–1 = 16 [−4 −12 24 ] = [ −1 −3 6 ]
4
8 8 −16 2 2 −4
0 1 −1 4 1 −2 0 2 3
1
Grammian matrix(A=M–1 AM) = 4 [−1 −3 6 ] [1 0 2 ] [8 2 1 ]
2 2 −4 1 −1 3 4 2 1
0 1 −1 0 2 3 4 0 0 1 0 0
1 1
= 4 [−1 −7 14 ] [8 2 1] = 4 [0 −12 4 ] = [0 3 1]
6 6 −12 4 2 1 0 0 12 0 0 3
1 0 0
∴J = M–1 AM = [0 3 1 ] Jordan Block
0 0 3

To compute state transition matrix,𝒆𝑨𝒕

The state transition matrix, 𝒆𝑨𝒕 = M Q(t) 𝒆𝚲𝒕 M–1

𝑡2 𝑡
1 𝑡 𝑒 𝜆1 0 0 𝑒𝑡 0 0
2 𝚲𝒕 𝑡
Q(t) = [0 1 𝑡] ; 𝒆 = [ 0 𝑒 𝜆2 0 ]=[0 𝑒 3𝑡 0]
0 0 1 0 0
𝑡
𝑒 𝜆3 0 0 𝑒 3𝑡

𝑡2
0 2 3 1 𝑡 𝑒𝑡 0 0 0 1 −1
1
𝑨𝒕 2
∴𝒆 = [8 2 1] [0 1 𝑡][0 𝑒 3𝑡 0 ] [ −1 −3 6] * 4
4 2 1 0 3𝑡 2 2 −4
0 1 0 0 𝑒

0 2 2𝑡 + 3 𝑒𝑡 0 0 0 0.25 −0.25
2
=[8 8𝑡 + 2 4𝑡 + 2𝑡 + 1] [ 0 𝑒 3𝑡 0 ] [ −0.25 −0.75 1.5 ]
4 4𝑡 + 2 2𝑡 2 + 2𝑡 + 1 0 0 𝑒 3𝑡 0.5 0.5 −1

10
M.Renuga / AP – ECE / AAMEC UNIT -5 CONTROL SYSTEM ANALYSIS USING STATE VARIABLE METHODS EC8391 – Control System engg

0 2𝑒 3𝑡 (2𝑡 + 3)𝑒 3𝑡 0 0.25 −0.25

=[8𝑒 𝑡 (8𝑡 + 2)𝑒 3𝑡 (4𝑡 2 + 2𝑡 + 1)𝑒 3𝑡 ] [−0.25 −0.75 1.5 ]
4𝑒 𝑡 (4𝑡 + 2)𝑒 3𝑡 (2𝑡 2 + 2𝑡 + 1)𝑒 3𝑡 0.5 0.5 −1

(𝑡 + 1)𝑒 3𝑡 𝑡𝑒 3𝑡 −2𝑡𝑒 3𝑡
=[(2𝑡 2 − 𝑡)𝑒 3𝑡 2𝑒 𝑡 + (2𝑡 2 − 5𝑡 − 1)𝑒 3𝑡 −2𝑒 𝑡 − (4𝑡 2 − 10𝑡 + 1)𝑒 3𝑡 ]
𝑡 2 𝑒 3𝑡 𝑒 𝑡 + (𝑡 2 − 2𝑡 − 1)𝑒 3𝑡 −𝑒 𝑡 − (2𝑡 2 − 4𝑡 − 1)𝑒 3𝑡

-------------------------------------------------------------------------------------------------------------------

CONCEPTS OF CONTROLLABILITY AND OBSERVABILITY

Controllability:
A system is said to be completely state controllable if it is possible to transfer the system state from any
initial state X(t0) to any other desired state X(td)in specified finite timemby a control vector U(t).\
Consider the state model of the system, 𝑋̇ = AX + BU ; Y = CX + DU …..(1)
1.Kalman’s test : 𝑸𝑪 = [𝑩 ∶ 𝑨𝑩 ∶ 𝑨𝟐 𝑩]
2.Gilbert’s test : The state model can be converted to canonical form by a transformation, X = MZ, where
M is the modal matrix and Z is the transformed state variable vector.
Eqn.(1): , 𝑍̇ ̃U
=𝐽Z+B ; Y = C̃Z + DU; ̃ = 𝑴−𝟏 𝑩 ; C̃ = 𝐶𝑀
where, 𝐽 = 𝑀−1 𝐴 𝑀 ; 𝐁
̃ must
Note: In this case the necessary and sufficient condition for complete controllability is that, the matrixB
have none of the rows with all zeros. If any row of the B̃ matrix is zero then the corresponding value is
uncontrollable.
Observability:
A system is said to be completely state observable if every state X(t) can be completely identified by
measurements of the output Y(t) over a finite time interval.
1.Kalman’s test : 𝑸𝒐 = [𝑪𝑻 ∶ 𝑨𝑻 𝑪𝑻 ∶ (𝑨𝑻 )𝟐 𝑪𝑻 ]
2.Gilbert’s test : ̃ = 𝑀−1 𝐵 ; ̃
𝐽 = 𝑀−1 𝐴 𝑀 ; B 𝐂 = 𝑪𝑴
𝒀(𝒔) 𝟑
1.A system is characterized by the transfer function 𝑼(𝒔) = .Identify the first state as the
𝒔𝟑 +𝟓𝒔𝟐 +𝟏𝟏𝒔+𝟔
output. Determine whether or not the system is completely controllable and observable.(May–13, 16
marks)
𝑌(𝑠) 3
Given : = ….(1)
𝑈(𝑠) 𝑠3 +5𝑠2 +11𝑠+6

Solution:
Cross multiplying the Eqn.(1):Y(s) (𝑠 3 + 5𝑠 2 + 11𝑠 + 6) = 3 U(s)
𝑑3 𝑦 𝑑2 𝑦 𝑑𝑦
The differential equation of the system is , +5 + 11 𝑑𝑡 + 6 𝑦 = 3 𝑢 ….(2)
𝑑𝑡 3 𝑑𝑡 2

𝑑3 𝑦 𝑑𝑦 𝑑2 𝑦
= − 6 𝑦 − 11 𝑑𝑡 − 5 +3𝑢 ….(3)
𝑑𝑡 3 𝑑𝑡 2

where y = output
𝑑𝑦 𝑑2 𝑦 𝑑3 𝑦
Let 𝑋1 = 𝑦 ; 𝑋1̇ = = 𝑋2 ; 𝑋2̇ = = 𝑋3 ; 𝑋3̇ = ….(4)
𝑑𝑡 𝑑𝑡 2 𝑑𝑡 3

Eqn.(4) in (3): 𝑋3̇ = − 6𝑋1 −11𝑋2 −5𝑋3 +3u

Hence the state model is,
𝑋1̇ 0 1 0 𝑋1 0 𝑋1
𝑋 ̇
[ 2] = [ 0 0 𝑋
1 ] [ 2 ] + [0 ] u ; y = [1 0 0 ] 𝑋
[ 2]
𝑋3 ̇ −6 −11 −5 𝑋3 3 𝑋3

𝑋̇ = AX + B U ––––––State Eqn. y = CX ––––––State Eqn

0 1 0 0
A=[ 0 0 1] ; B = [0] ; 𝐶 = [1 0 0]
−6 −11 −5 3

11
M.Renuga / AP – ECE / AAMEC UNIT -5 CONTROL SYSTEM ANALYSIS USING STATE VARIABLE METHODS EC8391 – Control System engg

KALMAN’S TEST FOR CONTROLLABILITY, 𝑸𝑪 = [𝑩 ∶ 𝑨𝑩 ∶ 𝑨𝟐 𝑩] ….(5)

0 0 1 0 0 0
B = [0] ; AB = [ 0 0 1] [0] = [ 3 ] ;
3 −6 −11 −5 3 − 15
0 1 0 0 3
𝟐
𝑨 𝑩 = A[𝑨𝑩] = [ 0 0 1 ][ 3 ] = [ −15 ]
−6 −11 −5 15 42
𝟎 𝟎 𝟑
Substitute the above terms in Eqn.(5): 𝑸𝑪 = [𝟎 𝟑 −𝟏𝟓] = – 27 ≠𝟎
𝟑 −𝟏𝟓 𝟒𝟐
Result: rank of 𝑄𝐶 is 3 , hence the system is completely controllable.
KALMAN’S TEST FOR OBSERVABILITY, 𝑸𝒐 = [𝑪𝑻 ∶ 𝑨𝑻 𝑪𝑻 ∶ (𝑨𝑻 )𝟐 𝑪𝑻 ] ….(6)
1 0 0 −6 0 0 −6 1 0
𝑻 𝑻 𝑻 𝑻
𝑪 =[ 0 ] ; 𝑨 = [1 0 −11] ; 𝑨 𝑪 = [1 0 −11] [ 0 ] = [ 1 ] ;
0 0 1 −5 0 1 −5 0 0

0 0 −6 0 0
𝑻 𝟐 𝑻 𝑻 𝑻 𝑻
(𝑨 ) 𝑪 = 𝑨 (𝑨 𝑪 ) = [1 0 −11] [ 1 ] = [ 0 ]
0 1 −5 0 1
𝟏 𝟎 𝟎
Substitute the above terms in Eqn.(6): 𝑸𝑶 = [𝟎 𝟏 𝟎] = 1 ≠𝟎
𝟎 𝟎 𝟏
Result: rank of 𝑄𝐶 is 3 , hence the system is completely observable.
––––––––––––––––––––––––––––––––––––––––
2. Determine the controllability and observability of the following system. (dec–12, 8 marks)

𝑋1̇ 0 1 0 𝑋1 0 𝑋1
[𝑋2̇ ] = [ 0 0 1 ] [𝑋2 ] + [ 0 ] u ; y = [1 0 0] [𝑋2 ]
𝑋3̇ 0 −2 −3 𝑋3 10 𝑋3

𝑋̇ = AX + B U ––––––State Eqn. y = C X ––––––State Eqn

0 1 0 0
A=[ 0 0 1] ; B=[0] ; 𝐶 = [1 0 0]
0 −2 −3 10

KALMAN’S TEST FOR CONTROLLABILITY, 𝑸𝑪 = [𝑩 ∶ 𝑨𝑩 ∶ 𝑨𝟐 𝑩] ….(5)

0 0 1 0 0 0
B=[0] ; 𝑨𝑩 = [ 0 0 1] [ 0 ] = [ 10 ] ;
10 0 −2 −3 10 −30
0 1 0 0 10
𝑨𝟐 𝑩 = A[𝑨𝑩] =[ 0 0 1] [ 10 ] =[ −30 ]
0 −2 −3 −30 70
𝟎 𝟎 𝟏𝟎
Substitute the above terms in Eqn.(5): 𝑸𝑪 = [ 𝟎 𝟏𝟎 −𝟑𝟎] = 1000 ≠𝟎
𝟏𝟎 −𝟑𝟎 𝟕𝟎
Result: rank of 𝑄𝐶 is 3 , hence the system is completely controllable.
KALMAN’S TEST FOR OBSERVABILITY, 𝑸𝒐 = [𝑪𝑻 ∶ 𝑨𝑻 𝑪𝑻 ∶ (𝑨𝑻 )𝟐 𝑪𝑻 ] ….(6)
1 0 0 0 0 0 0 1 0
𝑻 𝑻 𝑻 𝑻
𝑪 =[ 0 ] ; 𝑨 = [1 0 −2] ; 𝑨 𝑪 = [1 0 −2] [ 0 ] = [ 1 ] ;
0 0 1 −3 0 1 −3 0 0

0 0 0 0 0
𝑻 𝟐 𝑻 𝑻 𝑻 𝑻
(𝑨 ) 𝑪 = 𝑨 (𝑨 𝑪 ) = [1 0 −2] [ 1 ] = [ 0 ]
0 1 −3 0 1

12
M.Renuga / AP – ECE / AAMEC UNIT -5 CONTROL SYSTEM ANALYSIS USING STATE VARIABLE METHODS EC8391 – Control System engg

𝟏 𝟎 𝟎
Substitute the above terms in Eqn.(6): 𝑸𝑶 = [𝟎 𝟏 𝟎] = 1 ≠𝟎
𝟎 𝟎 𝟏
Result: rank of 𝑄𝐶 is 3 , hence the system is completely observable.
–––––––––––––––––––––––––––––––––––––––

3. Determine the state controllability and observability of the following system. (dec–10, 16
marks)

−3 1 1 0 1 𝑋1
̇ 0 0 1 𝑋
[𝑋] = [ −1 0 1] X + [0 0] u ; y = [ ] [ 2]
1 1 0
0 0 1 2 1 𝑋3
𝑋̇ = AX + B U ––––––State Eqn. y = CX ––––––State Eqn
−3 1 1 0 1
0 0 1
A = [ −1 0 1] ; B = [0 0] ; 𝐶=[ ]
1 1 0
0 0 1 2 1
KALMAN’S TEST FOR CONTROLLABILITY, 𝑸𝑪 = [𝑩 ∶ 𝑨𝑩 ∶ 𝑨𝟐 𝑩] ….(5)
0 1 −3 1 1 0 1 2 −2
B = [0 0 ] ; 𝑨𝑩 = [ −1 0 1] [0 0] = [2 0 ] ;
2 1 0 0 1 2 1 2 1
−3 1 1 2 −2 2 7
𝑨𝟐 𝑩 = A[𝑨𝑩] =[ −1 0 1] [2 0 ] = [0 3 ]
0 0 1 2 1 2 1
𝟎 𝟏 𝟐 −𝟐 −𝟐 𝟕
Substitute the above terms in Eqn.(5): 𝑸𝑪 = [𝟎 𝟎 𝟐 𝟎 𝟎 𝟑]
𝟐 𝟏 𝟐 𝟏 𝟐 𝟏
𝟏 𝟐 −𝟐
Consider , [𝟎 𝟐 𝟎 ] =6 ≠𝟎
𝟏 𝟐 𝟏
Result: rank of 𝑄𝐶 is 3 , hence the system is completely observable.
KALMAN’S TEST FOR OBSERVABILITY, 𝑸𝒐 = [𝑪𝑻 ∶ 𝑨𝑻 𝑪𝑻 ∶ (𝑨𝑻 )𝟐 𝑪𝑻 ] ….(6)
𝟎 𝟏 −𝟑 −𝟏 𝟎 𝟎 𝟏 𝟎 −𝟒
𝑪 𝑻 = [𝟎 𝟏] ; 𝑨𝑻 𝑪𝑻 = [ 𝟏 𝟎 𝟎 ] [𝟎 𝟏] =[𝟎 𝟏 ]
𝟏 𝟎 𝟏 𝟏 𝟏 𝟏 𝟎 𝟏 𝟐
−𝟑 −𝟏 𝟎 𝟎 −𝟒 𝟎 −𝟏𝟏
(𝑨𝑻 )𝟐 𝑪𝑻 = 𝑨𝑻 (𝑨𝑻 𝑪𝑻 ) = [𝟏 𝟎 𝟎 ] [ 𝟎 𝟏 ] = [𝟎 −𝟒 ]
𝟏 𝟏 𝟏 𝟏 𝟐 𝟏 −𝟏
𝟎 𝟏 𝟎 −𝟒 𝟎 𝟏𝟏
Substitute the above terms in Eqn.(6): 𝑸𝟎 = [𝟎 𝟏 𝟎 𝟏 𝟎 −𝟒]
𝟏 𝟎 𝟏 𝟐 𝟏 −𝟏
−𝟒 𝟎 𝟏𝟏
Consider, [ 𝟏 𝟎 −𝟒] = -5 ≠ 𝟎
𝟐 𝟏 −𝟏
Result: rank of 𝑄𝐶 is 3 , hence the system is completely observable.
Home work :
−𝟐 𝟏 𝟏
(i) A=[ ] ; B=[ ] ; C = [𝟏 −𝟏] ans: Controllable, not observable
𝟏 −𝟐 𝟎
0 0 0 40
(ii). A=[ 1 0 −3] ; B = [10] ; 𝐶 = [0 0 1]
0 1 −4 0
ans: Controllable, observable

13
M.Renuga / AP – ECE / AAMEC UNIT -5 CONTROL SYSTEM ANALYSIS USING STATE VARIABLE METHODS EC8391 – Control System engg

Gilbert’s test for controllability: ̃

𝐁=M–1 B
Evaluate the controllability of the system with
0 1 0 0
A=[ 0 0 1] ; B = [0] ; 𝐶 = [1 0 0]
−6 −11 −5 3
𝜆 −1 0
|𝜆𝐼 − 𝐴| = |0 𝜆 −1 | =0; i.e.,𝜆3 + 6𝜆2 + 11𝜆 + 6 = 0
6 11 𝜆+6
∴(𝜆 + 1) (𝜆 + 2) (𝜆 + 3) = 0; 𝜆1 = -1, 𝜆1 = -2, 𝜆1 = -3

Model matrix M is Vander monde matrix.

1 1 1 1 1 1
𝜆
M =[ 1 𝜆2 𝜆3 ] = [−1 −2 −3]
𝜆12 𝜆22 𝜆23 1 4 9

−6 6 −2 𝑇 −6 −5 −1
[−5 8 −3] [6 8 2] 3 2.5 0.5
𝐴𝑑𝑗[𝑀] −1 2 −1 −2 −3 −1
𝑀−1 = |𝑀|
= = = [−3 −4 −1]
−2 −2
1 1.5 0.5
3 2.5 0.5 0 0.5
̃ = M–1 B
𝐁 = [−3 −4 −1] [0] = [−1]
1 1.5 0.5 1 0.5
̃ , the system is completely controllable.
Result: As none of the element of 𝐁

--------------------------------------------------
Gilbert’s test for obsevrability: 𝐂̃ = CM

Evaluate the obsevrability of the system with

0 1 0 0
A = [0 0 1] ; B = [0] ; 𝐶 = [3 4 1]
0 −2 −3 1

𝜆 −1 0
|𝜆𝐼 − 𝐴| = |0 𝜆 −1 | =0; 𝜆3 + 3𝜆2 + 2𝜆 = 0; 𝜆 ( 𝜆2 + 3𝜆 + 2)
0 2 𝜆+3
𝜆1 = -1, 𝜆1 = -2, 𝜆1 = -3

Model matrix M is Vander monde matrix.

1 1 1 1 1 1
M = [𝜆1 𝜆2 𝜆3 ] = [0 −1 −2]
𝜆12 𝜆22 𝜆23 0 1 4

1 1 1
𝐂̃ = CM = [𝟑 𝟒 𝟏 ] [0 −1 −2] =[𝟑 𝟎 −𝟏]
0 1 4

Result: As there is one zero element in𝐂̃ , the system is not completely observable.

14
M.Renuga / AP – ECE / AAMEC UNIT -5 CONTROL SYSTEM ANALYSIS USING STATE VARIABLE METHODS EC8391 – Control System engg

15
M.Renuga / AP – ECE / AAMEC UNIT -5 CONTROL SYSTEM ANALYSIS USING STATE VARIABLE METHODS EC8391 – Control System engg

16
M.Renuga / AP – ECE / AAMEC UNIT -5 CONTROL SYSTEM ANALYSIS USING STATE VARIABLE METHODS EC8391 – Control System engg

17
M.Renuga / AP – ECE / AAMEC UNIT -5 CONTROL SYSTEM ANALYSIS USING STATE VARIABLE METHODS EC8391 – Control System engg

18
M.Renuga / AP – ECE / AAMEC UNIT -5 CONTROL SYSTEM ANALYSIS USING STATE VARIABLE METHODS EC8391 – Control System engg

19