MATLAB Premier

Asst. Prof. Dr. Melik DÖLEN

Middle East Technical University

Department of Mechanical Engineering

10/30/04 ME 304 1
 Outline
! Introduction
! Basic Features of MATLAB
! Prompt Level and Basic Aritmetic Operations
! Scalars, Vectors, and Matrices
! Matrix Operations and MATLAB Functions
! Graphics and Data Visualization
! Programming in MATLAB
! MATLAB in Control Engineering
! Operations on Transfer Functions
! Simulation of LTI Systems

10/30/04 ME 304 2
 Introduction
! MATrix LABoratory “MATLAB” is an interactive software for
numerical / symbolic computations and graphics.
! Originally designed for matrix computations, solving linear
equation sets, data visualization, etc.
! The capabilities of MATLAB can be extended through programs
written in its own programming language called “M-script”.
! Over the years, the capabilities of MATLAB in every aspect have
widened rapidly.
! It has now become an “industry-standard” computing
environment and an outstanding tool for engineering education.

10/30/04 ME 304 3
 MATLAB Prompt
To get started, type one of these: helpwin, helpdesk, or demo. For
product information, type tour or visit www.mathworks.com.
»
» (45 + 7)/3 Arithmetic operations: +, -, *, /, ^
ans = ans (ANSwer) is a temporary “variable”
17.3333
storing the result of an instant calculation.
» 100^2 – 3.5*ans Previous answer can be recalled.
ans =
1.0061e+004 Scientific notation: 1.0061×104
» sqrt(ans) MATLAB offers a large number of predefined
ans = functions like sqrt, sin, cos, tan, exp,
100.3029 log, atan, abs, tanh, and more...

10/30/04 ME 304 4
 MATLAB Objects
» a = 2.12 Result of various operations can be assigned to
a = the variables you want to define. Variable
2.1200 names are case sensitive.
» b = -9.75; Note that semicolon (;) supresses the display.

» x = [1; 2; 3] Column vector (3×1) is defined:

x = 1
1
2 x = 2
3
3

» y = [1 2 3] y is a row vector (1×3):

y =
1 2 3
y = [1 2 3]

10/30/04 ME 304 5
 MATLAB Objects (Cont’d)
» A = [1 2 3; 4 5 6; 7 8 9] A is a matrix (3×3):
A =
1 2 3 1 2 3
4 5 6 A = 4 5 6
7 8 9
7 8 9

» whos whos displays the names of all

Name Size Bytes Class variables defined in the MATLAB
“workspace” (memory).
A 3x3 72 double array
a 1x1 8 double array
ans 1x1 8 double array
b 1x1 8 double array
x 3x1 24 double array
y 1x3 24 double array

Grand total is 18 elements using 144 bytes

10/30/04 ME 304 6
 Special Variables
» i By default, i and j are defined as imaginary
ans =
0 + 1.0000i
unit (√-1).

» -5 + a*i MATLAB allows storage of complex numbers

ans = and relevant operations on them.
-5.0000 + 2.1200i

» cos(pi) cos(π)
ans =
-1
» a = inf; a = ∞ (see IEEE 754 standard!)
» 1/a
ans =
0

10/30/04 ME 304 7
 Important Notes
» sin(x) Most aritmetic operators and certain functions
ans = work for matrices and vectors as well.
0.8415
0.9093
0.1411

» help ans An extensive online help is available.

ANS Most recent answer.
ANS is the variable created automatically when expressions
are not assigned to anything else. ANSwer.

10/30/04 ME 304 8
 Array Manipulations
» A(2,3) Returns the element at 2nd row, 3rd column.
ans =
6
» A(:,2) Returns all the elements at 2nd column.
ans =
2
5
8 Column 1 2 3
» A(3,:) Returns all the
ans = elements at 3rd row.
7 8 9
1 2 3 Row 1
» A(1:2,2:3) Returns submatrix.
ans = A = 4 5 6 2
2 3
5 6 7 8 9 3

10/30/04 ME 304 9
 Array Manipulations (Cont’d)
» B = [A; zeros(1,3); y] Forms a new matrix B using the variables
B =
1 2 3
previously defined. Here, the command
4 5 6 zeros(m,n) creates an (m by n) matrix
7 8 9 consisting of 0 elements.
0 0 0
1 2 3

1 2 3
[A]3x3 4 5 6 Matrix A
B = [0]1x3 B = 7 8 9
[y]1x3
[ 0 0 0] Zero vector
[ 1 2 3] Vector y

10/30/04 ME 304 10
 Array Manipulations (Cont’d)
» B = [x ones(3,1) A] Redefines matrix B using the variables
B = previously defined. Here, the command
1 1 1 2 3
2 1 4 5 6
ones(m,n) creates an (m by n) matrix
3 1 7 8 9 consisting of 1s.

1 1 1 2 3
B = [x]3x1 [1]3x1 [A]3x3 B = 2 1 4 5 6
3 1 7 8 9

Vector x Unity Matrix A

vector

10/30/04 ME 304 11
 Matrix Operations
» A = [1 2 3; 4 5 6; 7 8 9]; Let us define these matrices.
» B = [1 2 0; 4 0 6; 0 8 9]
» C = [1 2; 3 4; 5 6];
» C’ Matrix inversion using (‘) operator.
ans = The command transpose(C)
1
2
3
4
5
6
serves for the same purpose.
» det(A) DETerminant of matrix A.
ans =
0

» inv(B) INVerse of matrix B.

ans =
0.4000 0.1500 -0.1000
0.3000 -0.0750 0.0500
-0.2667 0.0667 0.0667

10/30/04 ME 304 12
 Matrix Operations (Cont’d)
» A+B MATLAB enables you to perform
ans = arithmetic operations (+,-,*) on
2
8
4
5
3
12
matrices.
7 16 18

» A+C Matrices must be conformable to

??? Error using ==> + carry out a certain operation.
Matrix dimensions must agree.
» A*C Matrix multiplication.
ans =
22 28
49 64
76 100
» C*A Invalid operation due to a
??? Error using ==> * mismatch in matrix inner
Inner matrix dimensions must agree.
dimensions.

10/30/04 ME 304 13
 Matrix Operations (Cont’d)
» A-3 Scalars (1×1) are exceptions in
ans = matrix operations.
-2 -1 0
1 2 3
4 5 6

» 3*A Multiplies every element of the

ans = matrix A by 3.
3 6 9
12 15 18
21 24 27

» A.*B Multiplies the corresponding

ans = elements of matrix A and B.
1 4 0
16 0 36 The operator (.*) is called array
0 64 81 multiplication.

10/30/04 ME 304 14
 Solving Equation Sets
Consider this equation set: To implement this in MATLAB,
x1 + x 2 − 4 x 3 = 8 one defines the following:
2 x1 − 3x 2 + 2 x 3 = 1 » A = [1 1 -4; 2 -3 2; 5 -8 10];
5x1 − 8x 2 + 10 x 3 = 2 » b = [8; 1; 2];
» x = A\b
Alternatively, we have x =
5.5000
1 1 − 4  x1  8  3.5000
 2 − 3 2   x  = 1  ⇒ A⋅x = b 0.2500
  2   
5 − 8 10   x 3  2 The operator (\) is called left matrix
divide and it essentially solves the
Therefore, the solution becomes linear system using Gauss-Seidel
method. Note that x = inv(A)*b
( A −1A ) x = A −1b ⇒ x = A −1b
yields the same result.

10/30/04 ME 304 15
 Advanced Matrix Functions
» eig(A) Returns all the ! lu (command) computes LU
ans = factorization of a matrix.
eigenvalues of
5.0000
matrix A. ! chol computes the cholesky
4.0000
-1.0000
factorization of a symmetric
positive definite matrix.
» [V,D] = eig(A) ! qr computes the QR factorization
V = Eigenvectors of a matrix.
! svd obtains singular value
-0.7071 0.8058 -0.4082
-0.0000 0.0620 -0.8165
decomposition of a matrix.
0.7071 -0.5889 -0.4082 ! norm computes various matrix or
vector norms.
D = Eigenvalues ! cond, condest, rcond
5.0000 0 0 estimates various condition
0 4.0000 0 numbers.
0 0 -1.0000

10/30/04 ME 304 16
 Vectors
» t = 1:5 Creates a row vector whose components
t = increase arithmetically.
1 2 3 4 5
» t = 0:.1:1 Components can change by non-unit steps.
t =
Columns 1 through 7
0 0.1000 0.2000 0.3000 0.4000 0.5000 0.6000
Columns 8 through 11
0.7000 0.8000 0.9000 1.0000

» t = linspace(0,1,11) Generates a vector with 11 equally spaced

t = elements that lie between 0 and 1.
Columns 1 through 7
0 0.1000 0.2000 0.3000 0.4000 0.5000 0.6000
Columns 8 through 11
0.7000 0.8000 0.9000 1.0000

10/30/04 ME 304 17
 2D Graphics
! MATLAB offers a wide variety of
built-in data visualization functions.
! Function plot is used to generate
2D graphs.
! To obtain a graph of y=f(x), one
needs to create the domain x and
then form the range y with the
corresponding function values.

» x = linspace(0,2*pi,100);
» y = sin(x);
» plot(x,y);
» xlabel(‘x [rad]’);
» ylabel(‘y’);
» grid on;

10/30/04 ME 304 18
 2D Graphics (Cont’d)
As the second example, let
us plot the following function:
x
y = f (x) =
1+ x2

The corresponding MATLAB

commands are as follows:
» x = linspace(-5,5,100);
» y = x./(1+x.*x);
» plot(x,y);
» xlabel(‘x’);
» ylabel(‘y’);
» grid on;

10/30/04 ME 304 19
 Programming
! Entering all the commands at the MATLAB prompt sequentially
is a tedious and slow process.
! One can type in all the commands in a text file having an
extension of “m”. The file containing these commands is called
M-script.
! When the name of this “m” file is keyed in just like a command
at the prompt, MATLAB executes all the commands
sequentially.
! Similar to a high-level language, MATLAB provides a number of
standard programming constructs such as loops and
conditionals.

10/30/04 ME 304 20
 Conditionals
IF Statement ! The use of conditionals in
if <boolean expression 1>
MATLAB programs are very
MATLAB statements similar to those of common high-
elseif <boolean expression 2> level languages.
MATLAB statements
else ! The logical operators in MATLAB
MATLAB statements
end
are <, >, <=, >=, == (logical
equal), ~= (not equal).
Example ! Boolean expressions take the
if x == 0 values 1 (true) or 0 (false).
y = 0;
elseif x == 1
y = 1;
else
y = 1 – 2*(x – .25);
end

10/30/04 ME 304 21
 For Loop
FOR Statement ! for loop repeats the statements
for variable = expression as the loop index takes on the
MATLAB statements
end values in a given vector.
! Like an if construct, the loop
Example 1 Example 2 must be terminated by end
for k = 1:4 for k = [1 3 7 11] statement.
disp command simply displays
disp(k); disp(k);
end end !

its argument without showing the

variable name.
1 1
2 3
3 7
4 11

10/30/04 ME 304 22
 While Loop
WHILE Statement
! while loop repeats the statements as
long as the given boolean expression
while <boolean expression>
MATLAB statements
end is true.
! Another important command is break
Example which terminates the current for or
while loop.
x = 1;
while 1+x > 1
x = x/2;
end
! The sample program is used to test
disp(2*x) the floating point accuracy (“machine
epsilon”) of a particular computer
Output system.
2.2204e-016

10/30/04 ME 304 23
 MATLAB Functions
! While programming, it is often times necessary to define one’s own
subroutines (procedures/functions).
! Unlike high-level languages, these functions cannot be included in
the main program.
! Each function has to be an individual “m” file.
! This “m” file that begins with a line of the following form: function
[out1,out2,...] = cmd_name(in1,in2,...)
! When a function is invoked, MATLAB creates a temporary workspace.
The statements inside the function have no access to the variables
used in the main workspace unless they are passed as inputs.
! When execution ends, all local variables are erased.

10/30/04 ME 304 24
 MATLAB Functions (Cont’d)
function [y,z] = fcn(x) M-file named “fcn.m”.
z = 1 + x.*x;
y = x./z;
plot(x,y)

» x = linspace(-5,5,100);
» fcn(x) Calls M-file named “fcn.m” and
passes the vector x.
» y1 = fcn(x); Different output arguments were
» [y2,z2] = fcn(x);
passed from the MATLAB
function fcn.

10/30/04 ME 304 25
 Useful MATLAB Functions
Poles of a System Characteristic Polynomial
D(s) = s 5 + 3s 4 + 7s 3 + 5s 2 + 2s + 1 Let the poles of a system be
p1 = -3; p2 = -2;
» den = [1 3 7 5 2 1]; p3 = -1; p3,4= -5 ± 5i;
» roots(den)
ans =
-1.0777 + 1.9230i » poles = [-3 -2 -1 –5+5*i –5-5*i];
-1.0777 - 1.9230i » poly(poles)
-0.7742 ans =
-0.0352 + 0.5144i 1 16 121 416 610 300
-0.0352 - 0.5144i
The resulting characteristic polynomial is

D(s) = s 5 + 16s 4 + 121s 3 + 416s 2 + 610s + 300

10/30/04 ME 304 26
 Useful Commands (Cont’d)
Partial Fractions Expansion » num = [2 5 3 6];
» den = [1 6 11 6];
Consider the following TF: » [r,p,k] = residue(num,den)
r =
2s 3 + 5s 2 + 3s + 6 -6.0000
Residues
G (s ) = -4.0000
s 3 + 6s 2 + 11s + 6 3.0000

The resulting fractions are p =

-3.0000
N -2.0000 Poles
r
G (s ) = ∑ i + k -1.0000
i =1
s − pi
k =
−6 −4 3 2 Remainder
G (s ) = + + +2
s + 3 s + 2 s +1

10/30/04 ME 304 27
 Useful Commands (Cont’d)
Polynomial Multiplication Zero/Pole Locations
P(s) = s3 + 2s 2 + 3s + 1 » num = [2 5 3 6];
 D (s ) = P (s ) Q (s ) = ? » den = [1 6 11 6];
Q(s) = s + s + 5 
2
» pzmap(num,den)

» P = [1 2 3 1]; !"#$%&$'"() *+
856
» Q = [1 1 5];
» D = conv(P,Q) 8

D =
956
1 3 10 14 16 5

12 *3(-./0
9

The resulting polynomial becomes %956

%8
5 4 3 2
D(s) = s + 3s + 10s + 14s + 16s + 5
%856
%456 %4 %756 %7 %856 %8 %956 9
,$*#(-./0

10/30/04 ME 304 28
 Block Diagram Reduction
Transfer Functions in Series
» num1 = [1 5]; den1 = [1 3 5];
X(s) Y(s) » num2 = [1]; den2 = [1 1];
G1(s) G2(s) » [num,den] = series(num1,den1,num2,den2)
num =
0 0 1 5
Y (s )
= G (s) = G1 (s)G 2 (s)
X (s ) den =
1 4 8 5

As an illustration, let The resulting TF becomes

s+5 1 s+5 NUMerator
G1 (s) = ; G 2 (s ) = G (s ) =
s 2 + 3s + 5 s +1 s 3 + 4s 2 + 8s + 5 DENominator

10/30/04 ME 304 29
 BD Reduction (Cont’d)
Transfer Functions in Parallel
» num1 = [1 5]; den1 = [1 3 5];
G1(s) » num2 = [1]; den2 = [1 1];
» [num,den]=parallel(num1,den1,num2,den2)
X(s) + Y(s) num =
G2(s) +
0 2 9 10

den =
Y (s )
= G (s) = G1 (s) + G 2 (s) 1 4 8 5
X (s )

As an illustration, let The resulting TF becomes

s+5 1 2s 2 + 9s + 10
G1 (s) = ; G 2 (s ) = G (s ) =
s 2 + 3s + 5 s +1 s 3 + 4s 2 + 8s + 5

10/30/04 ME 304 30
 BD Reduction (Cont’d)
Closed Loop Transfer Functions
Y(s) » num1 = [1 5]; den1 = [1 3 5];
X(s) +
G1(s) » num2 = [1]; den2 = [1 1];
_ » [num,den]=feedback(num1,den1,num2,den2)
num =
G2(s) 0 1 6 5

den =
Y (s ) G1 (s)
= G (s) = 1 4 9 10
X (s ) 1 + G1 (s)G 2 (s)

As an illustration, let The resulting TF becomes

s+5 1 s 2 + 6s + 5
G1 (s) = ; G 2 (s ) = G (s ) =
s 2 + 3s + 5 s +1 s 3 + 4s 2 + 9s + 10

10/30/04 ME 304 31
 Unit Impulse Response
As an illustration, let

2s + 4
G (s ) =
s 2 + 2s + 4

MATLAB commands are

» num = [2 4]; den = [1 2 4];
» impulse(num,den,10)

Duration of simulation
(10 sec.)

10/30/04 ME 304 32
 Unit Step Response
As an illustration, let

2s + 4
G (s ) =
s 2 + 2s + 4

MATLAB commands are

» num = [2 4]; den = [1 2 4];
» step(num,den,10)

Duration of simulation
(10 sec.)

10/30/04 ME 304 33