EE301 Signals and Systems Analysis

Lecture 5: Classification of Signals (Cont.) and Classification of Systems

Credit: Some material in these slides is based on the references shown in the last slide.

Instructor: Dr. Abdulaziz A. Alorf

Department of Electrical Engineering

Qassim University

January 27th , 2025

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 1 / 31

Outline

1 Classification of Signals (Cont.)

2 Classification of Systems

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 2 / 31

Outline

1 Classification of Signals (Cont.)

2 Classification of Systems

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 3 / 31

Classification of Signals (Cont.)
Let’s continue the methods to classify signals:
Causal
5. CausalCausal Signals
Signals:Signals
- Causal signals
Causal are non-zero
Causal signals
signals only
are non-zero
are non-zero only
only fort tt⩾≥
forfor ≥000 (starts
(starts
(starts at t=
atattt =
= 0, 0,
0, or or later).
or later)
later)
2
2
Causal
Causal 1
1

-2 -1 0 1 2
-2 -1 0 1 2 tt

- Noncausal signalssignals
Noncausal are non-zero
are non-zero somett <
forforsome < 00 (starts
(startsbefore t =t0)= 0).
before
Noncausal signals are non-zero for some t < 0 (starts before t = 0)

Noncausal
Noncausal 1
1

-2 -1 0 1 2
-2 -1 0 1 2 tt
- Anticausal signals are non-zero only for t ⩽ 0 (goes backward in time
Anticausal signals
from t =Anticausal
0).
signals are
are non-zero
non-zero only for tt ≤
only for ≤ 00 (goes
(goes backward
backward in
in time
time
from t = 0)
from t = 0)
6. Real and imaginary signals: 2
Cuff (Lecture 1) 2 Systems
ELE 301: Signals and Fall 2011-12 39 / 45
Cuff (Lecture 1) ELE 301: Signals and Systems Fall 2011-12 39 / 45
Anticausal
Anticausal 1
EE301 Signals and Systems Analysis Lecture 51 January 27th , 2025 4 / 31
Classification of Signals (Cont.)

• So far, we mainly considered real (or integer) valued signals.

• Signals can also be complex as

z(t) = x(t) + jy(t) = r(t)ejϕ ,

√
where x(t) and y(t) are each real valued signals, j = −1, and r(t)ejϕ
represents the polar form representation of the signal.
• The conjugate of z(t) (i.e., z ∗ (t)) is sufficient to reproduce the parts of
z(t) such that:
∗
- Real part: x(t) = Re{z(t)} = z(t)+z 2
(t)
, where
z ∗ (t) = x(t) − jy(t) = r(t)e−jϕ is the conjugate of z(t).
∗
- Imaginary part: y(t) = Im{z(t)} = z(t)−z 2j
(t)
.
- Modulus (or p absolute value):
p
r(t) = |z(t)| = z(t)z ∗ (t) = x2 (t) + y 2 (t).  
y(t)
- Argument: ejϕ = ej arg z(t) =⇒ ϕ = arg z(t) = tan−1 x(t) .

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 5 / 31

 r is the modulus or magnitude of z
Classification of Signals (Cont.)
φ is the angle or phase of z
exp(jφ) = cos φ + j sin φ
Im Im
z = x + jy z = re jφ
r
φ

Re Re
• Complexcomplex
signalsexponential
are oftenofused
z = xto
+ represent
jy : waves and wavefunctions.
All such representations
z
e =e ultimately
x+jy
= e e = e (cos y +aj fundamental
x jydepend
x on sin y )
mathematical identity, known as Euler’s theorem, which takes the form
Know how to add, multiply, and divide complex numbers, and be able to
go between representations easily.
e±jϕ = cos(ϕ) ± j sin(ϕ).
Cuff (Lecture 1) ELE 301: Signals and Systems Fall 2011-12 42 / 45

From the previous point, the real and imaginary parts of this formula are
respectively given by
1 h ±jϕ
∗ i 1 h ±jϕ
∗ i
cos(ϕ) = e + e±jϕ , and sin(ϕ) = e − e±jϕ .
2 2j
EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 6 / 31
Classification of Signals (Cont.)
• A signal is said to be real when it satisfies the following condition:

x(t) = x∗ (t).

For example:
- x(t) = cos(ω0 t) is real, because x∗ (t) = cos(ω0 t).
- x(t) = at is real, because x∗ (t) = at.
• A signal is said to be imaginary when it satisfies the following condition:

x(t) = −x∗ (t).

For example:
- x(t) = jbt is imaginary, because

x∗ (t) = (jbt)∗ = −jbt → −x∗ (t) = jbt → −x∗ (t) = x(t).

• Consider the following complex signal:

x(t) = e−jω0 t = cos(ω0 t) − j sin(ω0 t) :

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 7 / 31
Classification of Signals (Cont.)

- This signal is complex (i.e., not real nor imaginary), but it is a mix of
two.
- To make the signal real then we need to set the imaginary part to zero
as
Set ω0 = π → x(t) = cos(πt) for some integer t.
- To make the signal imaginary then we need to set the real part to zero
as
π π
Set ω0 = → x(t) = −j sin( t) for some integer t.
2 2
- Therefore, we can conclude that:
a. For a real signal, the imaginary part must be zero.
b. For an imaginary signal, the real part must be zero.
• Complex signals arise naturally in many problems:
- Convenient representation for sinusoids.
- Communications.
- Radar, sonar, ultrasound.

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 8 / 31

Classification of Signals (Cont.)

• The complex exponential of z(t) = x(t) + jy(t) can be represented as:

ez(t) = ex(t)+jy(t) = ex(t) ejy(t) = ex(t) [cos(y(t)) + j sin(y(t))] .

• Please know how to add, multiply, and divide complex numbers, and be
able to go between representations easily.
7. Energy and power signals:
• Signal energy: A signal is said to be an energy signal when it has finite
energy. Energy of a continuous signal (complex or real) is given by
Z ∞
2
E= |x(t)| dt < ∞, if exists (in joules [J]).
−∞

Energy of a discrete signal (complex or real) is given by

∞
X
E= |x[n]|2 < ∞, if exists (in joules [J]).
n=−∞

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 9 / 31

Classification of Signals (Cont.)
• Signal power: A signal is said to be a power signal when it has finite
power. Power of a continuous signal (complex or real) is given by
Z T
1 2
2
P = 0 < lim |x(t)| dt < ∞, if exists (in watts [W]).
T →∞ T − T2

Power of a discrete signal (complex or real) is given by

N
1 X
P = 0 < lim |x[n]|2 < ∞, if exists (in watts [W]).
N →∞ 2N + 1
n=−N

• An energy signal requires: 0 < E < ∞, and P → 0.

- This means that when you integrate a nonzero finite power signal over
all time gives you energy.
• A power signal requires: 0 < P < ∞, and E → ∞.
- This means that when you integrate a nonzero finite power signal over
infinite time, you get infinite energy.
EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 10 / 31
Classification of Signals (Cont.)
• Examples:
(1) Is the following cosine signal x(t) a power or energy signal? And then
find its power or energy?

x(t) = A cos(ω0 t + ϕ)

- Checking if x(t) is a power signal:

Z T
1 2
P = lim A2 cos2 (ω0 t + ϕ)dt
T →∞ T − T2
T
A2
Z
1 2
= lim [1 + cos(2ω0 t + 2ϕ)]dt,
T →∞ T 2 − T2
1 1
Note: cos2 (x) = + cos(2x)
2 2
: 0
 
! 

 A2 T 2 Z T2  A2

A  
= lim · + 0 
cos(2ω
 t + 2ϕ)dt = .
T
T →∞  2 2T


− T2 
 2

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 11 / 31
Classification of Signals (Cont.)

∴ x(t) is a power signal.

- Checking if x(t) is an energy signal:
Z ∞
E= A2 cos2 (ω0 t + ϕ)dt → ∞
−∞

∴ x(t) is not an energy signal.

(2) Is the following signal x(t) a power or energy signal? And then find
its power or energy?

x(t) = e−at u(t), where a > 0

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 12 / 31

Classification of Signals (Cont.)

- Checking if x(t) is a power signal:

Z T
1 2
P = lim e−2at u2 (t)dt
T →∞ T − T2
Z T
1 2
= lim e−2at dt
T →∞ T 0
"
0 1!#
1 1 −∞
*
 0
= lim e  −e
 ,
−2a

T →∞ T
Z b  b
±kx 1 ±kx
Note: e dx = e
a ±k a
1
= lim → 0.
T →∞ 2aT

∴ x(t) is not a power signal.

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 13 / 31

Classification of Signals (Cont.)

- Checking if x(t) is an energy signal:

Z ∞
E= e−2at u2 (t)dt
−∞
Z ∞
= e−2at dt
0
" 1!#
1 −∞ *0 
 0
= e  −e
−2a


1
= .
2a
∴ x(t) is an energy signal.

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 14 / 31

Classification of Signals (Cont.)

• A very handy property: If the energy of the signal x(t) is given by E,

then the energy of the time scaled signal x(at) is Ea , where a is positive.
To show that
Z +∞
2 r
|x(at)| dt, substitute for r = at 7→ t =
−∞ a
Z +∞
2 dr
= |x(r)|
−∞ a
Z +∞
1 2
= |x(r)| dr
a −∞
| {z }
=E
E
= joules.
a

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 15 / 31

Classification of Signals (Cont.)
• Some signals are neither power nor energy signals because they have
unbounded power and energy, i.e., E and P → ∞. For example, the
following signal x(t) is neither power nor energy:

x(t) = eat u(t), where a > 0.

Exercise: Try to prove it at home ©.

• Some remarks:
- All finite duration signals with finite amplitude are energy signals.
- All periodic signals are power signals, but the reverse is not always
right! When we determine that a signal is a power signal from its
periodicity, then we can use the following simplified equation to
compute its power:
2
A2

A
P = √ = , where A is the signal amplitude.
2 2
- The continuous impulse is neither an energy nor a power signal
(E and P → ∞), because its amplitude goes to infinity (unbounded).
But the discrete impulse has energy of 1 joule, and 0 watt.
EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 16 / 31
Outline

1 Classification of Signals (Cont.)

2 Classification of Systems

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 17 / 31

Classification of Systems

There are many ways to classify systems:

1. Linear and nonlinear systems:
• A system S is said to be linear if and only if it satisfies each one of the
following two conditions:
(1) Homogeneity: If x is any signal and a is any scalar then

S(ax) = a S(x),

which means that scaling before or after the system is the same.
(2) Superposition: If x and x
e are any two signals then

S(x + x
e) = S(x) + S(e
x),

which means that summation before or after the system is the same.
• Linearity means that the following pairs of block diagrams are
equivalent, i.e., have the same output for any input(s):

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 18 / 31

 Linearity means the following pairs of block diagrams are equivalent, i.e.,
Classification
have the sameof Systems
output for any input(s)

x1
x y S
y
a F +

x2

x1
S
x y y
F a +
x2
S

Superposition
• Examples: Systems with the following responses y(t), check whether
Examples of nonlinear systems: sign detector, multiplier (sometimes),
each system is linear or not?
comparator, quantizer, adaptive filter
(1) A buffer system: y(t) = x(t), this system is linear because
- S(ax) = ax & aS(x) ELE
Cuff (Lecture 2)
= 301: → S(ax)
ax Signals = aS(x), and Fall 2011-12 51 / 70
and Systems
- S(x + x e) = x + xe & S(x) + S(e x) = x + x e → S(x + xe) = S(x) + S(e x).
2
(2) A multiplier as a squaring system: y(t) = x (t), this system is
nonlinear because
Multiplier as a modulator, y (t) = x(t) cos(2πft), is linear.
EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 19 / 31
Classification
MultiplierofasSystems
a modulator, y (t) = x(t) cos(2πft), is linear.
x + y

- S(ax) = a2 x2 & aS(x) = ax2cos(2π

→ S(ax) f t) ̸= aS(x), and
2
- S(x + x
e) = (x + x e) & S(x) + S(e x) = x22 + x e2 →
Multiplier as a squaring
e) ̸= S(x) + S(e
S(x + x x). system, y (t) = x (t) is nonlinear.

x + y

Multiplier as a Squaring System

(3) Multiplier as a modulator: y(t) = x(t) cos(2πf t), this system is linear
because
- S(ax) = ax cos(2πf t) & aS(x) = ax cos(2πf t) → S(ax) = aS(x), and
- S(x
Cuff+ e) =2) cos(2πf t) ×ELE
x
(Lecture +x
(x 301: ) =and
eSignals x cos(2πf
Systems t) + x
e cos(2πf t) & 52 / 70
Fall 2011-12
S(x) + S(e x) = x cos(2πf t) + x e cos(2πf t) → S(x + x e) = S(x) + S(e x).

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 20 / 31

Classification of Systems

x + y

cos(2π ft)
Multiplier as a Modulator
2. Time variant and time invariant systems:
• A system is said to be time invariant if its input and output
characteristics do not change y
x with time.+ Mathematically, a system is
time-invariant if a time shift in the input produces the same time shift in
the output. For a system S:

y(t) = Sx(t) (continuous-time representation),

implies that
y(t − τ ) = Sx(t − τ ),

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 21 / 31

Classification of Systems

for any time shift τ . Similarly,

y[n] = Sx[n] (discrete-time representation),

implies that
y[n − k] = Sx[n − k],
for any number of samples k.
Implies that delay and the system F commute. These block diagrams
• Previous definitions implies that the delay and the system S commute.
are equivalent:
The following block diagrams are equivalent:
x y
Delay by τ S

x y
S Delay by τ

Time invariance is an important

EE301 Signals and Systems Analysis
system property. ItJanuary
Lecture 5
greatly simplifies
27th , 2025 22 / 31
Classification of Systems
• Examples: Systems with the following responses y[n], check whether each
system is time invariant or not?
(1) y[n] = x[n] + x[n − 2], this system is time invariant because
- y[n − k] = x[n − k] + x[n − k − 2], and
- Sx[n − k] = x[n − k] + x[n − k − 2],
- then y[n − k] = Sx[n − k].
(2) y[n] = x[n] + nx[n − 3], this system is time variant because
- y[n − k] = x[n − k] + (n − k)x[n − k − 3], and
- Sx[n − k] = x[n − k] + nx[n − k − 3],
- then y[n − k] ̸= Sx[n − k].
• Finally, time invariance is an important system property. It greatly
simplifies the analysis of systems.
3. Linear time variant (LTV) and linear time invariant (LTI) systems:
• A system is said to be LTV when it satisfies both linearity and time
variance.
• A system is said to be LTI when it satisfies both linearity and time
invariance.
EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 23 / 31
Classification of Systems
• Examples: Check whether each one of the following systems is LTV or
LTI?
(1) y[n] = nx2 [n], this system is not LTV nor LTI because it is not
linear as shown below:
- S(ax) = na2 x2 & aS(x) = nax2 → S(ax) ̸= aS(x), and
- S(x + xe) = n(x + x e)2 & S(x) + S(ex) = nx2 + ne
x2 →
S(x + xe) ̸= S(x) + S(ex).
(2) y[n] = nx[n], this system is an LTV system because it is linear and
time variant as shown below:
• Checking linearity:
- S(ax) = nax & aS(x) = nax → S(ax) = aS(x), and
- S(x + xe) = n(x + x e) = nx + nex & S(x) + S(e
x) = nx + ne x→
S(x + xe) = S(x) + S(e x),
- therefore, the system is linear.
• Checking time variance/invariance:
- y[n − k] = (n − k)x[n − k], and
- Sx[n − k] = nx[n − k],
EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 24 / 31
Classification of Systems

- then y[n − k] ̸= Sx[n − k],

- therefore, the system is time variant.
(3) y(t) = x(t − t0 ), this system is an LTI system. Please try to prove
it at home ,.
4. Static and dynamic systems:
• A system is memoryless if the output depends only on the present input,
such as an ideal amplifier, an ideal gear, lever in a mechanical system
etc. The following systems are static:

y[n] = x[n], and y(t) = 10x2 (t).

• A system with memory has an output signal that depends on inputs in

the past or future, such as energy storage circuit elements (capacitors
and inductors), and springs or moving masses in mechanical systems.
The following systems are dynamic:

y[n] = x[n] + x[n − 2], and y(t) = x(t) + x(t + 5).

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 25 / 31

Classification of Systems
5. Causal and noncausal systems:
• A causal system has an output/response that depends ONLY on past
input, present input, or both. A noncausal system’s output depends on
the future inputs.
- All noncausal systems are dynamic systems.
- Not all dynamic systems are noncausal.
- Causal systems can be static or not.
- All static systems are causal.
6. Invertible and non-invertible systems:
• A system is invertible if the input signal can be recovered from the
output signal. If the system S is an invertible system, and

y = Sx

then there is an inverse system S IN V such that

x = S IN V y = S IN V Sx

so S IN V S = I, the identity operator.

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 26 / 31
 so F INV F = I , the identity operator.
Classification of Systems
x y x
S SINV

- Consider an LTI system with the input signal x(t), then we are getting
the output/response y(t) of the system according to its impulse
Cuff (Lecture 2) ELE 301: Signals and Systems Fall 2011-12 58 / 70
response h(t). This process can be represented in time domain using
the convolution theorem such that

y(t) = h(t) ∗ x(t).

- To find a system inverse, we need to find the inverse of h(t). It is much

simpler to find the inverse in the frequency domain. Then let’s Fourier
transform the previous equation as follows:

ℑ{y(t)} = ℑ{h(t) ∗ x(t)} = H(f )X(f ) = Y (f ).

Therefore, a system inverse can be obtained by computing H −1 (f ),

and then performing the inverse Fourier transform as:

ℑ−1 {Y (f )} = ℑ−1 {
H(f −1
 (f ) X(f )} ⇐⇒ y(t) = x(t).
) H 
 
EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 27 / 31
Classification of Systems

7. Stable and unstable systems:

• A system is said to be stable if it produces a bounded output for a
bounded input. If a bounded input

|x(t)| ≤ Mx < ∞,

always results in a bounded output

|y(t)| ≤ My < ∞,

where M x and M y are finite positive numbers, the system is Bounded

Input Bounded Output (BIBO) stable.
• The best test signal used to test any system is the unit step function.
That’s because it is a bounded signal, where we know that its value is 1
at any time from 0 to ∞.
• Examples: Check whether each one of the following systems is stable or
not?

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 28 / 31

Classification of Systems

3
(1) y[n] = (100 x[n]) , this system is stable because its response/output
is bounded when the system’s input is a step function such that

y[n] = (100 u[n])3 = 1000, 000 u[n].

R∞
(2) y(t) = −∞ x(t)dt, this system is not stable because its response is a
ramp signal which is unbounded as shown below
Z ∞ Z ∞ Z t
y(t) = x(t)dt = u(t)dt = u(t)dt = t, t ≥ 0 = r(t).
−∞ −∞ 0

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 29 / 31

The End

Any questions?

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 30 / 31

References

B.P. Lathi. Modern Digital and Analog Communication Systems, 3rd ed. Oxford
University Press, Oxford, 2009.
Sanjay Sharma. Signals and Systems, 6th ed. S.K. Kataria & Sons, Delhi, 2008.

Alan V. Oppenheim, Alan S. Willsky, and S. Hamid Nawab. Signals and

Systems, 2nd ed. Pearson, London, 1996.
Simon Haykin and Barry Van Veen. Signals and Systems, 2nd ed. Wiley,
Hoboken, 2002.
Michael D. Adams. Signals and Systems, 3rd ed. University of Victoria,
Victoria, 2020.

EE301 Signals and Systems Analysis Lecture 5 January 27th , 2025 31 / 31