PSG COLLEGE OF TECHNOLOGY, COIMBATORE 641 004.

DEPARTMENT OF INSTRUMENTATION AND CONTROL SYSTEMS ENGINEERING

19U502 – CONTROL SYSTEMS II

DIAGNOLISATION USING SIMILARITY TRANSFORMATION

22U210 – BOOMISHA S

22U230 – POORNIMA DEVI M

22U234 – RITHIKA SRI A

22U254 – THARUNIGHA S
14-10-2024 PSG COLLEGE OF TECHNOLOGY 1
EIGEN VALUES AND EIGEN VECTORS
A non-zero column vector 'v' is an eigen vector of a square matrix A if there exists a scalar λ such that
Av = λv
Where λ is Eigen value or characteristic value of A.

The eigenvalues of an n*n matrix A are the roots of the characteristic equation
| λI - A | = 0
This equation may be expressed in expanded form

The values of λ are called as eigen values of A matrix and the equation is called as characteristic equation of A
The above characteristic equation yields n eigen values λi where i= 0,1,2,…............n

14-10-2024 PSG COLLEGE OF TECHNOLOGY 2

• On comparing |λI - A | = 0 and |sI - A | = 0 we can conclude that eigenvalues of state model and poles of
system transfer function are same .
• System's stability depends on location of eigenvalues of state model. Thus the system is stable if the
eigenvalues have negative real parts.

From Av = λ v i i i

(λiI - A)vi = 0

Let v = mi be the solution of this above equation. The solution mi is called eigenvector of A associated
with the eigenvalue λ . i

The solution of above equation depend upon the rank r of the matrix (λI - A).
If r is the rank of the matrix then there are (n-r) independent solutions (eigen vectors).

14-10-2024 PSG COLLEGE OF TECHNOLOGY 3

If the eigenvalues of the matrix A are all distinct then we have only one independent eigenvector
corresponding to any particular eigen value λ . i

The eigenvector can be obtained by taking cofactors of the matrix (λ I - A) along any row that is
i

Mi = where k = 1,2,3,4.........n

Where Cki are the cofactors of the matrix (λ I - A).

i

APPLICATIONS OF EIGEN VALUES:

Diagonalization of Matrix Solving systems of linear equation Stability analysis

Data compression Principle component analysis Signal Processing

14-10-2024 PSG COLLEGE OF TECHNOLOGY 4

14-10-2024 PSG COLLEGE OF TECHNOLOGY 5
 SIMILARITY TRANSFORMATION
• A similarity transformation in a control system is a mathematical operation that changes the
representation of a system while preserving its essential characteristics.
• It is a process to simplify a matrix while preserving its essential properties.
Formula: B=P-1 AP , where
A - Original matrix
P - Invertible matrix
P-1 - Inverse of matrix P
B - Transformed matrix (simpler form)
NEED OF SIMILARITY TRANSFORMATION :
• Simplifies Eigenvalue Computation
• Matrix Simplification
• Preserves Properties
• Stability and Control Systems

14-10-2024 PSG COLLEGE OF TECHNOLOGY 6

PROPERTIES OF SIMILARITY TRANSFORMATION

• Invariance of eigenvalues

• Invariance of determinant

• Invariance of trace

• Preservation of system structure

• Transforming state-space representations

• Invariance of stability

• Decoupling systems

• Reversibility

• Invariance of controllability and observability

14-10-2024 PSG COLLEGE OF TECHNOLOGY 7

PROOF FOR INVARIANCE OF DETERMINANT

Let A and B are similar matrices and P be the transformation matrix which transforms A to B by a similarity
transformation, P-1AP = B .
∴ B = P-1AP ....(1)
On taking determinant of equation(1) we get,
|B| = |P-1AP| .....(2)
Since the determinant of a product of two or more square matrices is equal to the product of their individual
determinants, the equation(2) can be written as,
|B| = |P-1| |A| |P| = |A| |P-1| |P|
= |A| |P-1P| = |A| |I| (∵P-1P=1)
= |A| (∵|I|= 1)
From the above analysis it is evident that the determinant of a matrix is invariant under a similarity transformation.
14-10-2024 PSG COLLEGE OF TECHNOLOGY 8
PROOF FOR INVARIANCE OF CHARACTERISTIC EQUATION AND EIGENVALUES

Let A and B are similar matrices and P be the transformation matrix which transforms A to B by a similarity
transformation,P-1 AP = B
The characteristic equation of matrix B is given by
|λI – Β| = 0 ......(3)
On substituting B=P-1 AP in equation(3) we get,
|λI – Β| = |λI – P-1 AP |
= [λP-1P - P-1AP | (∵ P-1P = 1)
= [P-1 (λΙ – Α)P | .....(4)
Since the determinant of a product is the product of the determinant, the equation(4) can be written as,
|λ1 - B| = |P-1| |λ1 - A| |P| = |λI-A| |P-1| |P|
= |λ1-A| |P-1 P|
= |λ1-A| |I| (∵ P-1P=1)
= |λ1 - A| (∵ |I|= 1)
From the above analysis it is clear that the characteristic equations of A and B are identical. Since the characteristic
equations are identical, the eigenvalues of A and B are identical. Hence the eigenvalues are invariant under a similarity
(linearity) transformation.
14-10-2024 PSG COLLEGE OF TECHNOLOGY 9
PROOF FOR INVARIANCE OF TRACE OF A MATRIX

Let A and B are similar matrices and P be the transformation matrix which transforms A to B by a similarity

transformation, P-1AP = B.

∴ tr B = tr P-1AP .....(5)
For an n x m matrix C and m x n matrix D, regardless of whether CD = DC or
CD ≠ DC, we have,
tr (CD) = tr (DC) ....(6)
Using the property of equation(6), the equation(5) can be written as,

tr B = tr AP P-1

= tr AI = tr A ( ∵ PP-1 = I and AI = A)
From the above analysis it is clear that the trace of a matrix is invariant under a similarity transformation.

14-10-2024 PSG COLLEGE OF TECHNOLOGY 10

 DIAGONALIZATION
If a nxn matrix has n independent eigenvectors, we will be able to
perform a similarity transformation , to obtain a matrix that has the
eigenvalues of original matrix on the diagonal.

14-10-2024 PSG COLLEGE OF TECHNOLOGY 11

14-10-2024 PSG COLLEGE OF TECHNOLOGY 12
14-10-2024 PSG COLLEGE OF TECHNOLOGY 13
14-10-2024 PSG COLLEGE OF TECHNOLOGY 14
14-10-2024 PSG COLLEGE OF TECHNOLOGY 15
14-10-2024 PSG COLLEGE OF TECHNOLOGY 16
14-10-2024 PSG COLLEGE OF TECHNOLOGY 17
14-10-2024 PSG COLLEGE OF TECHNOLOGY 18
14-10-2024 PSG COLLEGE OF TECHNOLOGY 19
14-10-2024 PSG COLLEGE OF TECHNOLOGY 20
14-10-2024 PSG COLLEGE OF TECHNOLOGY 21
MATRIX A IN PHASE VARIABLE CANONICAL FORM WITH DISTINCT EIGEN
VALUES
The n*n matrix A will be in the form of

The transformation x = Pz will be given as will transform P -1AP into the diaganol matrix of
form

14-10-2024 PSG COLLEGE OF TECHNOLOGY 22

Problem
The state space model given:

It can be written in a standard form as

Where,

the eigenvalues by solving the characteristic equation

det⁡(λI - A)=0
(λ+1) (λ+2) (λ+3)=0

λ1 = -1 λ2 = -2 λ3 = -3
14-10-2024 PSG COLLEGE OF TECHNOLOGY 23
Three eigen values found are distinct
λ1 = -1 λ2 = -2 λ3 = -3

P-1AP =

Diagonal matrix:

14-10-2024 PSG COLLEGE OF TECHNOLOGY 24

MATRIX A IN PHASE VARIABLE CANONICAL FORM WITH REPEATED EIGEN
VALUES
The n*n matrix A will be in the form of

The transformation x = Sz will be given as will transform S -1AS into the diagnol matrix of form

14-10-2024 PSG COLLEGE OF TECHNOLOGY 25

Problem:

0 1 0
Matrix A: 0 0 1
-4 -9 -6

the eigenvalues by solving the characteristic equation det⁡(λI - A)=0

det(λI - A)=0 (λ+1) 2 (λ+4)=0

The eigenvalues are:

λ1 = -1 (multiplicity 2) λ2 = -4 (multiplicity 1)

1 0 1 1 0 1
S = λ1 1 λ3 = -1 1 -4
λ12 2λ1 λ32 1 2 16

14-10-2024 PSG COLLEGE OF TECHNOLOGY 26

REPEATED EIGEN VALUES:

1.14 0.095 -0.047 0 1 0 1 0 1

0.571 0.714 0.142 0 0 1 -1 1 -4
S−1 AS = -0.142 -0.095 0.047 -4 -9 -6 1 2 16

-1 1 0
0 -1 0
0 0 -4
=

This is in the form of Jordan canonical form

14-10-2024 PSG COLLEGE OF TECHNOLOGY 27

APPLICATIONS:
• Solving Linear State-Space Systems: Diagonalization transforms the state equation into decoupled first-
order systems. Easier to compute the state transition matrix and solve state-space equations.
• Modal Analysis in Control Systems: Stability depends on the location of eigenvalues in the complex
plane.
Diagonalization makes it easier to analyze the eigenvalues for stability.
• Controllability and Observability Analysis: Diagonalization helps check whether the system is
controllable or observable. Eigenvalues give insight into which modes are controllable/observable.
• Designing State Feedback Controllers: Diagonalization simplifies pole placement for desired system
behavior. Allows control of individual system modes.

14-10-2024 PSG COLLEGE OF TECHNOLOGY 28

REFERENCES:

• Ogata K , "Modern Control Engineering", Prentice-Hall of India Pvt Ltd, New Delhi, 2010.
• Nagrath I J, Gopal M , "Control System Engineering", 5th Edition, New Age International Pvt Ltd,
2008.
• Nagoor Kani, "Advanced Control Theory" , 2nd Edition,

14-10-2024 PSG COLLEGE OF TECHNOLOGY 29