Basics of control engineering

Mathematical tools
-
Simon CHESNE

1
Outline -Mathematical tools
• Elementary functions
• Convolution integral
• Frequency domain representations
• Fourier transform
• Laplace transform

2
 Elementary functions

The Heaviside step function, or the unit

step function, usually denoted by H or θ
(but sometimes u, 1 or 𝟙), is a
discontinuous function, named after Oliver
Heaviside (1850–1925), whose value is
zero for negative arguments and one for
positive arguments.

3
Elementary functions

The ramp function is functions whose

graph is shaped like a ramp.
Its derivative is the Heaviside function

4
Elementary functions

The rectangular pulse function is also

called the rectangle function, boxcar
function, or gate function.

5
 Elementary functions
The Dirac delta distribution

The Dirac delta can be loosely thought of as a

function on the real line which is zero everywhere
except at the origin, where it is infinite.
 pour t = 0
 (t ) = 
 0 pour t  0
and which is also constrained to satisfy the identity
+

  (t )dt = 1
−

du (t )
We can consider (t) as the derivative of the step function  (t ) =
dt 6
 Elementary functions
The Dirac delta distribution

+

  (t )dt = 1
−

 pour t = 0
 (t ) = 
 0 pour t  0
7
 Convolution integral
• The convolution of x and h is written x∗h, denoting the operator with the symbol ∗.
It is defined as the integral of the product of the two functions after one is
reversed and shifted. As such, it is a particular kind of integral transform:


x(t )  h(t ) = x(t )* h(t ) =  x( )h(t −  )d
−

8
 Convolution integral
1. Express each function in terms of a dummy
variable
2. Reflect one of the functions
3. Add a time-offset, t, which allows to slide along
the axis.
4. Start t at −∞ and slide it all the way to +∞.
Wherever the two functions intersect, find the
integral of their product.
The resulting waveform (not shown here) is the
convolution of functions f and g.

9
Convolution integral

10
Convolution integral
• Properties:
• Dirac – Impulse response
• Temporal response of a system

11
Convolution integral
• Dirac – Impulse response
x(t)

M y(t)

If f(t) is a unit impulse, the result of this process is simply h(t).

𝛿 𝑡 ∗ ℎ 𝑡 = ℎ(𝑡) 𝛿 𝑡 − 𝑎 ∗ ℎ 𝑡 = ℎ(𝑡 − 𝑎)
12
Convolution integral
Input System Output

h(t) y(t)
x(t)

Time domain

x(t)
H(ω)

M y(t)
Frequency domain
k
Convolution integral
Single impulse response h(t).

Each « impulse » will

contribute to the response.

e(t)
eΔt (t)
𝑒 𝑡 ≈ ෍ 𝑒∆𝑡 (𝑡) ≈ ෍ 𝑒𝑖 𝛿(𝑡 − 𝑖∆𝑡)

e(2Δt) Then the system response can be written:

e(Δt)

𝑦 𝑡 ≈ ෍ 𝑒𝑖 𝛿 𝑡 − 𝑖∆𝑡 ∗ ℎ(𝑡)
Δt t
𝑦 𝑡 = 𝑒(𝑡) ∗ ℎ(𝑡)
The input can be written as the sum of impulse
15
16
17
Convolution integral
• Practical example
• Measurement of the impulse response
Output
System (h(t) the impulse response)
Input (Impulse)

𝛿 𝑡 ∗ ℎ 𝑡 = ℎ(𝑡)

• Now we can simulate the response y(t) for all forms of input e(t).

𝑒(𝑡) ∗ ℎ(𝑡) = 𝑦 𝑡 18
Exercises and Examples
• Cruise control model
• Rotational motion: Satellite Attitude control model

19
Frequency domain representations
• Time domain Vs frequency domain

From time domain to…

… frequency domain

20
Frequency domain representations
+∞

• Fourier transform 𝑋(𝑗𝜔) = න 𝑥 𝑡 𝑒 −𝑗𝜔𝑡 𝑑𝑡

−∞

+∞

• Laplace transform 𝑋(𝑠) = න 𝑥 𝑡 𝑒 −𝑠𝑡 𝑑𝑡

0

Fourier is used primarily for steady state signal analysis, while

Laplace is used for transient signal analysis.

In the follow up, we will usually employ a lower case letter for a time-
domain signal, and an upper case letter for Laplace-domain signals

21
Frequency domain representations
• Properties
• Convolution of time-domain signals

where * is the convolution operator

• Product of time-domain signals

22
Frequency domain representations
• Properties
• Integration theorem:

• Derivation theorem

23
Frequency domain representations
• Laplace transform of common signals
• Dirac

“Proof”

24
Frequency domain representations
• Laplace transform of common signals
• Step function

“Proof”

25
Frequency domain representations
• Inverse Laplace transform
• Complex calculus

• Not used!
• In practice, we prefer breaking down the function X(s) into a sum of simple
elements, then use the inverse Laplace transform table and the linearity
property of this operator for finding the corresponding time-domain signal.

26
Frequency domain representations

27
Frequency domain representations

28
Frequency domain representations
• Classical functions (t>0)

29
Frequency domain representations
• Using Laplace transform to solve differential equations

u(t) y(t)
System

30
Exercice – introduction to block diagrams
• Solve this system using Simulink
• Input: d and e (step functions)
• Output: a

31