Chapter 1 SIGNALS and SYSTEMS: MATHEMATICAL PRELIMINARIES

What is a signal?
The concept of signal refers to the space or time variations in the physical state of an object. The quantification of this
state is used to represent, store or transmit a message. The course is not about the message but about the properties of
the signal.

Continuous-time and discrete-time signals

Examples and mathematical representation

Signals are met in diverse fields, such as:

Elec. Eng. e.g. voltages/currents in a circuit
speech signals
image signals
Physics e.g. radiation
Mech. Eng. e.g. vibration studies
Astronomy e.g. (2-D) pulsars, distant stars
Biomedicine e.g. EEG, ECG, retinoscopy, MRI
Seismology e.g. tectonic plate movement, earthquake prediction
Economics e.g. level of trading in stock market

Signals are represented mathematically as functions of one or more independent variables. In this course we focus our
attention on signals involving a single independent variable. For convenience, we will usually refer to this variable as
time.

We will be considering two basic types of signals.

• Continuous-time signals. In this case the independent variable is continuous, i.e., these signals are defined for a
continuum of values of the independent variable.
• Discrete-time signals. In this case the independent variable takes a discrete set of values, i.e., these signals are
defined only at discrete times.

Observed signal x(t)

10

-5

-10
0 10 20 30 40 50 60
continuous time
Observed signal x[n]
10

-5

-10
0 10 20 30 40 50 60
discrete time
In the figure above you can see a continuous and a discrete sine wave described by the following equations,
respectively.
• Continuous case: x (t ) = A cos(ω 0 t + φ )
• Discrete case: x[n ] = A cos(θ 0 n + ψ )

Elementary transformations of the independent variable

In this section we focus on a class of elementary signal transformations that involve simple modifications of the
independent variable, i.e., the time axis. These transformations allow us to introduce several basic properties of signals
and systems. We consider the following.
• Time shift. Starting from the signal x(t ) in continuous time, time shift refers to the operation that gives us either
the signal x (t − t 0 ), t0 > 0 that is a delayed version of x(t ) or the signal x(t + t 0 ) that is an advanced version of
x(t ) . For discrete time signals we would have x[n], x[n − n 0 ], x[ n + n 0 ] respectively.
• Time reversal. Starting from the signal x(t ) in continuous time, time reversal refers to the operation that gives us
the signal x(−t ) that is a reflection of x(t ) about t = 0 . For discrete time signals we would have x[n] and x[−n]
respectively.
• Time scaling. Starting from the signal x(t ) in continuous time, time scaling refers to the operation that gives us
the signal x(αt ) that is (i) linearly stressed if α < 1 , (ii) linearly compressed if α > 1 and (iii) reversed in time if
α < 0 . For discrete time signals we would have x[n] and x[αn] respectively.

Fundamental classes of signals

Deterministic and random signals

The following classification can be made in real life signals.

• If the value of a signal at some specific time is determined exactly by its model then we have a deterministic signal.
• If the value of the signal cannot be determined but only some statistical properties of it, e.g., probability of
occurrence of a certain value at some specific time, then we have a random or stochastic signal.

R and o m signal x (t )
1

0 .5

-0 .5

-1
0 20 40 60 80 100
R and o m b inary signal x[n ] tim e
1

0 .5

-0 .5

-1
0 20 40 60 80 100

1
Periodic and non-periodic signals

• For a continuous signal, if we can find a number T0 such that x(t + T0 ) = x(t ) for all t , then the signal is periodic.
In the case of a discrete signal we have x[n + N 0 ] = x[n ] for all n .
• If there are no such numbers T0 or N 0 the signal is non-periodic.

P eriodic signal x(t)

4

-2

-4
0 10 20 30 40 50 60 70 80

time
P eriodic signal x[n]
4

-2

-4
0 10 20 30 40 50 60 70 80

Even and odd signals

• For a continuous signal, if x ( −t ) = x (t ) for all t , then the signal is even. In the case of a discrete signal we have
x[− n] = x[ n] for all n .
• For a continuous signal, if x(−t ) = − x(t ) for all t , then the signal is odd. In the case of a discrete signal we have
x[− n] = − x[ n] for all n . An odd signal is 0 at t = 0 or n = 0 , since x(0) = − x(0) ⇒ x(0) = 0 .
Any signal can be broken into a sum of two signals, one of which is even and one of which is odd. This is because
1 1 1 1 1 1
x ( t ) = x ( t ) + x ( − t ) − x ( −t ) = x ( t ) + x ( t ) + x ( − t ) − x ( −t ) =
2 2 2 2 2 2
1 1 1 1 1 1
= x (t ) + x ( −t ) + x (t ) − x ( −t ) = [ x (t ) + x ( −t )] + [ x (t ) − x ( −t )]
2 2 2 2 2 2
1 1
where [ x (t ) + x ( −t )] is even signal and [ x (t ) − x ( −t )] is odd signal.
2 2

Causal signal models

If x(t ) = 0, for t < t 0 or for discrete signals x[n] = 0, for n < n 0 then we have causal signals. The starting point t 0 or n 0
is very often taken to be the origin.

Special signal models

The sinusoid

• Continuous case x(t ) = A cos(ω0t + φ )

2
• Discrete case x[n] = A cos(θ 0 n + ψ )
ω0
where A is the amplitude, ω 0 is the radian frequency, θ 0 = is the normalised frequency and φ ,ψ are the phase
2π
shifts.

Observed signal x(t )

10

-5

-10
0 10 20 30 40 50 60
continuous time
Observed signal x[n]
10

-5

-10
0 10 20 30 40 50 60
discrete time

The causal real exponential signal

• Continuous case x (t ) = Ae −αt , t > 0 . An example is shown in the figure below with α > 0 .
• Discrete case x[n] = Bλ n , n ≥ 0 . An example is shown in the figure below with λ real and λ < 1 .

E x p o n e n tia l s ig n a l x (t) E x p o n e n tia l s ig n a l x [n ]

1 1

0 .9 0 .9

0 .8 0 .8

0 .7 0 .7

0 .6 0 .6

0 .5 0 .5

0 .4 0 .4

0 .3 0 .3

0 .2 0 .2

0 .1 0 .1

0 0
5 10 15 20 5 10 15 20
tim e

3
The general complex exponential signal

• Continuous case x (t ) = Aλαt where A and α are, in general complex numbers and λ is real.
• Discrete case x[n ] = Aλαn with A , α and λ as in the continuous case.
A very important class of complex exponentials is obtained by constraining α to be purely imaginary. Specifically
2π
consider the signal x (t ) = e jω0t . This is periodic with period T = . Throughout this course we will deal with this
ω0
type of signal very frequently.

The delta function or unit impulse function δ (t )

This is a signal that is everywhere equal to zero except at the origin. This is defined as
⎧ 0 t≠0
δ (t ) = ⎨
⎩+ ∞ t = 0
It is conveniently defined as assigning the value x (τ ) to the signal x (t ) at the instant t = τ through the integral
+∞ +∞
∫ x(t )δ (t − τ )dt = ∫ x(t + τ )δ (t )dt =x(τ )
−∞ −∞
In discrete time the equivalent function is defined as
⎧0 n ≠ 0
δ [ n] = ⎨
⎩1 n = 0
and the following relationships hold.
n = +∞ n = +∞
∑ x[n ]δ [n − k ] = ∑ x[n + k ]δ [n ] = x[k ]
n = −∞ n = −∞

An important property
+∞ +∞
The relationship ∫ δ (t − τ )dt = ∫ δ (t )dt =1 holds.
−∞ −∞
Thus δ (t ) may be interpreted as the pulse (not unique) in the limit as ε → 0 (see following figure).

δ (t )

1
AREA=1 ε

−ε ε t
2 2

Comment
Observe that other functions have this property. For example consider the function
1 α
x (t ) = ( 2 2 )
π α +t
+∞
α +∞ 1 α 1 t
with α real. The area under the curve is ∫ dt = ( tan −1 ) = 1.
π −∞ α + t
2 2
π α α −∞
This is independent of α . Hence as this parameter tends to zero we have a delta function.

4
There are many other functions we encounter for instance in Systems Theory, Communications etc., such as
⎧ 1 2 ⎫
e −t / ε ⎬
2
• δ (t ) = lim ⎨
ε →0 ε π
⎩ ⎭
⎧ sin(t / ε ) ⎫
• δ (t ) = lim ⎨ ⎬
ε →0 ⎩ πt ⎭

The unit step function u(t )

This is defined as
⎧0 t < 0
u(t ) = ⎨
⎩1 t ≥ 0
The unit step function can be related to the unit impulse function as follows.
t
u(t ) = ∫ δ (τ )dτ
−∞
Replacing τ with t − σ we have τ → −∞, σ → +∞ , τ = t , σ = 0 and dτ = −dσ so that
0 +∞
u(t ) = ∫ δ (t − σ )( −dσ ) = ∫ δ (t − σ )dσ
+∞ 0
du(t )
δ (t ) = u ′(t ) =
dt
In discrete time the equivalent function is defined as
⎧0 n < 0
u[n ] = ⎨
⎩1 n ≥ 0
and the following relationships hold.
n +∞
u[n ] = ∑ δ [k ] = ∑ δ [n − k ]
k = −∞ k =0
δ [n ] = u[n ] − u[n − 1]

Signal energy and power

Periodic Signals

We define the power of a periodic signal to be the mean square of its values over one period, i.e.,
1 T /2 2
E= ∫ x(t ) dt
T −T / 2
Thus, if for example x(t ) represents a voltage signal, E represents the power dissipated in a resistor of 1Ω.
Examples
1 T /2 2 2 2π
1. x (t ) = A sin(ω 0 t + φ ) then E = ∫ A sin (ω 0 t + φ )dt where T = ω or equivalently
T −T / 2 0

A2 T /2
A2
E=
2T
∫ [1 − cos(2ω 0 t + 2φ )]dt = 2
−T / 2
T /2
∫ [A1 cos(ω 0 t + φ1 ) + A2 cos(2ω 0 t + φ 2 )] dt or
1
2. x(t ) = A1 cos(ω 0 t + φ1 ) + A2 cos(2ω 0 t + φ 2 ) then E = 2
T −T / 2

1 T /2 2 A2 + A22
E= ∫ [ A1 cos 2 (ω 0 t + φ1 ) + 2 A1 A2 cos(ω 0 t + φ1 ) cos(2ω 0 t + φ 2 ) + A22 cos 2 ( 2ω 0 t + φ 2 )]dt = 1
T −T / 2 2
Question: What would the answer be when
3. x (t ) = A1 cos(ω1t + φ1 ) + A2 cos(ω 2 t + φ 2 ) with ω 1 and ω 2 unrelated or
4. x(t ) = A1 cos(ω 0 t + φ1 ) + A2 cos(2ω 0 t + φ 2 ) + … + An cos(nω 0 t + φ n )

In the discrete time the energy of a periodic signal may be defined as

5
 1
∑ x[n ]
2
E=
N samples
within
a period

Non-periodic signals

It is more convenient in this case to define the signal energy as

+∞
2
E = ∫ x (t ) dt
−∞

⎧⎪e −αt , t ≥ 0
For example if x(t ) = ⎨
⎪⎩0, t < 0
then
+∞
1
E = ∫ e −2αt dt =
0 2α
Signals for which E is finite are called "energy signals", or, in mathematical terms square-integrable.
Also the energy expended over the time interval t1 ≤ t ≤ t 2 may be defined as
t2
2
E = ∫ x (t ) dt
t1
In the discrete time the total energy of a non-periodic signal may be defined as
n = +∞
E = ∑ x[n ]
2

n = −∞
or the energy expended over the time interval n1 ≤ n ≤ n 2 as
n = n2
E = ∑ x[ n ]
2

n = n1

Continuous-time and discrete-time systems

Definition of system

A system operates on a signal (the input) to modify, transform or re-express it in another form (the output) which may
be more desirable.

Example of system
The input voltage E (t ) produces an output voltage V 0 (t ) .

C
E (t ) V 0 (t )

input output

A continuous-time system is a system in which continuous-time input signals are applied and result in continuous-time
output signals. A discrete-time system is a system in which discrete-time input signals are applied and result in discrete-
time output signals.

Interconnection of systems

• Series or cascade interconnection. The output of System 1 is the input to System 2.

System 1 System 2
Input 6 Output
• Parallel interconnection. The same input signal is applied to Systems 1 and 2.

System 1

Input + Output

System 2

• Combination of both cascade and parallel interconnection.

System 1 System 2

System 4
Input + Output

System 3

• Feedback interconnection. The output of System 2 is fed back and added to the external input to produce the
actual input to System 1.

Input + System 1 Output

System 2

Basic system properties

Systems with and without memory

The following classification can be made.

• Memoryless systems. The output at instant t for continuous-time systems or at instant n for discrete-time systems
depends only on the value of input at the same instant t or n . For example y[n] = αx[n] or y (t ) = αx(t ) .
• Systems with memory. The output at instant t for continuous-time systems or at instant n for discrete-time
systems depends not only on the value of input at the same instant t or n , but also on past or future values of the
input. For example y[n] = αx[n] + βx[ n − 1] .

Invertible and non-invertible systems

The following classification can be made.

• Invertible systems. A system is called invertible, if another system exists that is called inverse that, when cascaded
with the original system, yields an output equal to the input to the first system.

Examples
1. Consider the memoryless system described by the input-output relationship y[n ] = αx[n ], α ≠ 0 . This is
1
invertible with inverse x[n ] = y[n ] .
α

7
 n
2. Consider the system with memory y[n] = ∑ x[k ] , which is called accumulator or summer. This is invertible
k = −∞
n n −1
with inverse x[n] = y[n] − y[n − 1] . This is because y[n ] = ∑ x[k ] = ∑ x[k ] + x[n ] ⇒ x[n ] = y[n ] − y[n − 1] .
k = −∞ k = −∞
• Non-invertible systems. A system is called non-invertible, if no system exists that, when cascaded with the original
system, yields an output equal to the input to the first system.
Examples
1. Consider the memoryless system described by the input-output relationship y[n] = 0 . This is non-invertible.
2. Consider the memoryless system y[n] = x 2 [ n] . This is non-invertible, since from knowledge of the output we
cannot determine the sign of the input.

Obviously a system is invertible if distinct inputs lead to distinct outputs.

Causal and non-causal systems

The following classification can be made.

• Causal or non-anticipative systems. The output at instant t for continuous-time systems or at instant n for
discrete-time systems depends on the value of input at the same instant t or n and on past values of the input but
not on future values of the input. For example y[n] = αx[n] + βx[ n − 1] .
• Non-causal or anticipative systems. The output at instant t for continuous-time systems or at instant n for
discrete-time systems depends on future values of the input. For example y[n] = αx[ n] + βx[n + 1] .

Stable and unstable systems

• Stable systems. A stable system is one in which all bounded inputs lead to bounded outputs. For example consider
the system y[n] = e x[ n ] . Let B be an arbitrary positive number, and let x[n] an arbitrary signal bounded by B ,
i.e., x[ n] < B or − B < x[n] < B for all n . In that case e − B < e x[ n ] < e B , i.e., it is guaranteed that if the input is
bounded by a positive number B , the output is bounded by a positive number e B . Thus, the system is stable.
• Unstable systems. An unstable system is one in which not all bounded inputs lead to bounded outputs. For example
n
consider the system y[n] = ∑ u[ k ] = ( n + 1)u[ n] , with u[n ] the unit step function. In that case the input u[n] = 1
k = −∞
but the output y[n] increases without bound as n increases.

Time-invariant and time-varying systems

• Time-invariant (TI) systems. A TI system is one in which if y[n] is the output when the input x[n] is applied, then
y[n − n 0 ] is the output when x[ n − n 0 ] is applied. For continuous-case we would have the signals
y (t ), x(t ), y (t − t 0 ), x(t − t 0 ) respectively. Conceptually, a system is TI if the behaviour and the characteristics of
the system are fixed over time. For example the system y[n] = αx[ n] .
• Time-varying systems. A time-varying system is one in which if y[n] is the output when the input x[n] is applied,
then y[n − n 0 ] is not necessarily the output when x[n − n 0 ] is applied. For example consider the system
y[n] = nx[n] . In that case if y1 [n] is the output to the input x1 [n] , then the output to the input x 2 [n ] = x1 [n − n0 ]
is y 2 [n] = nx 2 [n] = nx1 [ n − n 0 ] ≠ ( n − n 0 ) x1 [n − n 0 ] = y1 [ n − n 0 ] . Thus, the system is not time-invariant.

Linear and non-linear systems

• Linear systems. A linear system is one that possesses the property of superposition, i.e., if the input consists of the
weighted sum of several signals, then the output is the weighted sum of the responses of the system to each of
those signals. Thus, a linear system possesses the following two properties.
1. Additivity property. If the response to x1 (t ) is y1 (t ) and the response to x 2 (t ) is y 2 (t ) then the response to
the signal x1 (t ) + x 2 (t ) is y1 (t ) + y 2 (t ) .
2. Homogeneity property. If the response to x1 (t ) is y1 (t ) , then the response to the signal αx1 (t ) is αy1 (t ) .
• Non-linear systems. At least one of the above properties does not hold.