Lecture 1 Examples

Signals and Systems Course

Example: 1. Energy and Power of a Causal Decaying Exponential Signal Problem: Determine the
total energy and the average power of the signal:

x(t) = Ae−at u(t)

where A is a real constant and a > 0.

Causal Decaying Exponential Signal: x(t) = Ae−at u(t)

x(t)
A

Ae−1

1
a t

Figure 1: The signal x(t) is a causal decaying exponential that starts at a value of A at t = 0 and decays
towards zero as t → ∞.

Solution:
Total Energy:
∞ ∞ ∞
e−2at A2
Z Z 
2 2 −2at 2
E= |x(t)| dt = A e dt = A =
0 0 −2a 0 2a

Average Power (Improved Clarity):

T T
A2 1 − e−2aT
Z Z  
1 1
Pav = lim |x(t)|2 dt = lim A2 e−2at dt = lim =0
T →∞ 2T −T T →∞ 2T 0 T →∞ 2T 2a

A2
Answer: E = 2a , Pav = 0

Example: 2. Average Power of Complex Exponential and Sinusoidal Signals Problem: Determine the
average power of:
1. x1 (t) = Aej(ω0 t+θ)
2. x2 (t) = A cos(ω0 t + θ)

Solution:
Part 1: |x1 (t)|2 = |Aej(ω0 t+θ) |2 = |A|2 . The average power is constant, so Pav,1 = |A|2 .

1
 A2
Part 2: |x2 (t)|2 = A2 cos2 (ω0 t + θ) = 2 (1 + cos(2ω0 t + 2θ))
T0
A2 A2
Z
1
Pav,2 = (1 + cos(2ω0 t + 2θ))dt =
T0 0 2 2
2
A
Answer: Pav,1 = |A|2 , Pav,2 = 2

Example: 3. Time-Shifting of Continuous and Discrete-Time Signals Problem: Given signals:

(
A(t − 1) for 1 ≤ t ≤ 2
x(t) = (1)
0 otherwise
x[n] = δ[n − 1] + δ[n − 2] − δ[n − 3] (2)

Find: yadv (t) = x(t + 1), ydel (t) = x(t − 1), yadv [n] = x[n + 1], ydel [n] = x[n − 1].
(
A(t − 1), 1≤t<2 Discrete-Time Signal: x[n] = δ[n − 1] + δ[n − 2] − δ[n − 3]
Continuous-Time Signal: x(t) =
0, otherwise
x[n]
x(t) 1 1
1
A

n
−1 1 2 3 4 5

−1
t −1
1 2

(a) The original CT signal x(t). (b) The original DT signal x[n].

Figure 2: The original signals for time-shifting.

Solution:
CT Signal Shifting:
(
A(t − 2) for 2 ≤ t ≤ 3
ydel (t) = x(t − 1) = (3)
0 otherwise
(
At for 0 ≤ t ≤ 1
yadv (t) = x(t + 1) = (4)
0 otherwise

DT Signal Shifting:

ydel [n] = x[n − 1] = δ[n − 2] + δ[n − 3] − δ[n − 4] (5)

yadv [n] = x[n + 1] = δ[n] + δ[n − 1] − δ[n − 2] (6)

Example: 4. Combined Time-Scaling and Time-Shifting Problem: Given:


1
 for − 1 < t < 0
x(t) = 1 − 0.5t for 0 ≤ t < 2

0 otherwise


Find y(t) = x(−3t − 4).

2
 
1,
 −1 ≤ t ≤ 0
Continuous-Time Signal x(t) = 1 − 0.5t, 0<t≤2

0, otherwise


x(t)

1
1

1−
0.5
t

t
−1 2

Figure 3: The original CT signal x(t).

Solution:

Method 1: Algebraic Substitution (Most Reliable)

Substituting τ = −3t − 4: Interval 1: −1 < τ < 0 =⇒ −1 < −3t − 4 < 0 =⇒ − 34 < t < −1,
where y(t) = 1. Interval 2: 0 ≤ τ < 2 =⇒ 0 ≤ −3t − 4 < 2 =⇒ −2 < t ≤ − 43 , where
y(t) = 1 − 0.5(−3t − 4) = 3 + 1.5t.

Method 2: Graphical Step-by-Step (Correct Order)

For the general form y(t) = x(at + b), it’s crucial to factor the expression inside the parentheses to isolate
the time shift:
4
y(t) = x(−3t − 4) = x(−3(t + ))
3
This form clearly shows the sequence of operations: **SCALE FIRST, THEN SHIFT**.
Step 1: Time Scaling and Reversal Start with x(t) and apply the scaling factor −3. This
involves:

1. Scaling: Compress the signal by a factor of 3 (replace t with 3t)

2. Reversal: Reflect the signal about the vertical axis (replace t with −t)
This gives us the intermediate signal v(t) = x(−3t).

3
 Step 1: Time Scaling and Reversal by -3

x(−3t)
x(t) (Original)
x(−3t) (Corrected)

t
−1 − 23 − 13 1 2 2
3 3

Figure 4: Step 1: Time scaling and reversal x(−3t) - compress by 3 and reflect.

4
Step 2: Time Shift Now shift the intermediate signal v(t) = x(−3t) left by 3 units to get:

4 4
y(t) = v(t + ) = x(−3(t + )) = x(−3t − 4)
3 3

4
Step 2: Time Shift Left by 3

x(−3(t + 4/3))
x(−3t) (Correct Starting Signal)
x(−3t − 4) (Correct Final Result)

t
−2 − 43 −1 − 32 1
3

4
Figure 5: Step 2: Time shift left by 3 to get final result x(−3t − 4).

Common Mistakes in Graphical Method

Many students make these errors:

ˆ Wrong Order: Applying shift before scaling, which gives incorrect results

ˆ Incorrect Factoring: Not properly factoring x(at + b) as x(a(t + ab ))

ˆ Sign Confusion: Confusing left shift vs. right shift when dealing with negative coefficients

ˆ Incorrect Intermediate Steps: Not creating the proper intermediate signal v(t) = x(at) before
shifting

4
Verification
Both methods should yield the same result:

1
 for − 43 < t < −1
y(t) = 3 + 1.5t for − 2 < t ≤ − 43

0 otherwise


Final Transformed Signal y(t) = x(−3t − 4)

y(t)

1
1
t
1.5
3+

t
−2 − 43 −1

Figure 6: The final transformed signal y(t) = x(−3t − 4) obtained by both methods.

Key Takeaway
For combined transformations x(at + b):
1. Always factor first: x(at + b) = x(a(t + ab ))
2. Apply scaling/reversal first: Create intermediate signal v(t) = x(at)

3. Then apply shift: y(t) = v(t + ab ) = x(a(t + ab ))

The algebraic substitution method is more reliable and less prone to errors than the graphical method.

5
 Lecture 2 Examples
Signals and Systems Course

Example: [ 1. Fundamental Periodicity of a Continuous-Time Sinusoid] Problem: Show that x(t) =

A cos(ω0 t + θ) is periodic with fundamental period T0 = 2π/ω0 .

Sinusoidal Signal: x(t) = A cos(ω0 t + ϕ)

x(t)
A

t
t0 t0 + T0

2π
T0 = ω0
−A

Figure 1: A generic plot of the continuous-time sinusoidal signal x(t) = A cos(ω0 t + θ), illustrating its
periodicity.

Solution:
For periodicity: x(t + T0 ) = x(t)

x(t + T0 ) = A cos(ω0 (t + T0 ) + θ) = A cos(ω0 t + ω0 T0 + θ)

For x(t + T0 ) = x(t): ω0 T0 = 2πk for integer k
The smallest positive T0 occurs when k = 1:
2π
T0 =
ω0
2π
Answer: T0 = ω0

Example: [ 2. Periodicity of the Sum of Continuous-Time Signals] Problem: Given:

x1 (t) = sin(10πt) (1)

x2 (t) = sin(20πt) (2)
x3 (t) = sin(31t) (3)

Determine if x4 (t) = x1 (t) + x2 (t) and x5 (t) = x1 (t) + x3 (t) are periodic. If periodic, find the
fundamental period.
Solution:

1
 Fundamental periods:
2π 1
T1 = = (4)
10π 5
2π 1
T2 = = (5)
20π 10
2π
T3 = (6)
31
For x4 (t) = x1 (t) + x2 (t):
T1 1/5
= =2
T2 1/10
Since the ratio is rational, x4 (t) is periodic with T4 = 15 .
For x5 (t) = x1 (t) + x3 (t):
T1 1/5 31
= =
T3 2π/31 10π
Since the ratio is irrational, x5 (t) is not periodic.

Sum of Sinusoids and Fundamental Period

Signal Amplitude
2 x1 (t) = sin(10πt)
x2 (t) = sin(20πt)
x4 (t) = x1 (t) + x2 (t)
1

t
−T4 T4 2T4

−1

−2
1
Fundamental Period T4 = 5

Figure 2: A plot of x4 (t). We can see that x1 (t) (with period T1 = 1/5) completes one cycle while x2 (t)
(with period T2 = 1/10) completes two cycles. The sum signal x4 (t) repeats every T4 = 1/5 seconds.

Answer: x4 (t) is periodic with T4 = 51 ; x5 (t) is not periodic.

Example: [ 3. Even and Odd Decomposition of a Triangular Pulse] Problem: Given the continuous-time
signal x(t) shown below, find its even component, xe (t), and its odd component, xo (t).
The signal x(t) is a triangular pulse defined mathematically as:

t
 for 0 ≤ t ≤ 1
x(t) = 2 − t for 1 < t ≤ 2

0 otherwise


2
 
t,
 0≤t≤1
Triangular Pulse Signal x(t) = 2 − t, 1<t≤2

0, otherwise


x(t)
1

t
1 2

Figure 3: The original triangular pulse signal x(t).

Solution:
Any signal can be decomposed into even and odd components:
1 1
xe (t) = [x(t) + x(−t)], xo (t) = [x(t) − x(−t)]
2 2
Step 1: Find x(−t) Reflect x(t) about t = 0: x(−t) exists for t ∈ [−2, 0].

Time Reversal (Reflection) of a Triangular Pulse

x(t)
1 x(t) (Original)
x(−t) (Reversed)

t
−2 −1 1 2

Figure 4: The time-reversed signal x(−t), obtained by reflecting x(t) about the vertical axis.

1 1
Step 2: Calculate xe (t) For t > 0: xe (t) = 2 x(t) (since x(−t) = 0) For t < 0: xe (t) = 2 x(−t)
(since x(t) = 0) At t = 0: xe (0) = 0

3
 Even Component: xe (t) = 21 [x(t) + x(−t)]

Signal Amplitude
1 x(t)
x(−t)
xe (t)

1
2

t
−2 −1 1 2

Figure 5: The even component xe (t) = 12 [x(t) + x(−t)]. Note its symmetry about the vertical axis.

Step 3: Calculate xo (t) For t > 0: xo (t) = 21 x(t) For t < 0: xo (t) = − 21 x(−t) At t = 0: xo (0) = 0

Odd Component: xo (t) = 12 [x(t) − x(−t)]

Signal Amplitude
1
x(t)
−x(−t)
1
2 xo (t)

t
−2 −1 1 2

− 12

−1

Figure 6: The odd component xo (t) = 21 [x(t) − x(−t)]. Note its anti-symmetry about the vertical axis.

Verification: xe (t) + xo (t) = x(t) for all t.

Answer: xe (t) = 12 [x(t) + x(−t)], xo (t) = 12 [x(t) − x(−t)]

Example: [ 4. Properties of a Finite-Duration Discrete-Time Signal] Problem: Consider the discrete-

time signal x[n] defined as: (
A, n = 0, 1, 2
x[n] =
0, otherwise
where A is a real constant.
Determine the following:
1. The even and odd parts of x[n], denoted as xev [n] and xod [n].

2. The total energy E and average power Pav of x[n].

4
 Discrete-Time Rectangular Pulse: x[n] = A(u[n] − u[n − 3])

x[n]
A A A
A

n
−3 −2 −1 1 2 3 4

Figure 7: The discrete-time signal x[n] is non-zero only at three points, n = 0, 1, 2, where it has a
constant amplitude A.

Solution:
Part 1: Even and Odd Parts
1 1
xev [n] = (x[n] + x[−n]), xod [n] = (x[n] − x[−n])
2 2
For x[n] = A at n = 0, 1, 2 and x[−n] = A at n = 0, −1, −2:
n x[n] x[−n] xev [n] xod [n]
-2 0 A A/2 −A/2
-1 0 A A/2 −A/2
0 A A A 0
1 A 0 A/2 A/2
2 A 0 A/2 A/2

A,
 n=0
xev [n] = A/2, n = ±1, ±2

0, otherwise


A/2,
 n = 1, 2
xod [n] = −A/2, n = −1, −2

0, otherwise


Discrete-Time Even Component: xev [n] = 12 [x[n] + x[−n]] Discrete-Time Odd Component: xod [n] = 21 [x[n] − x[−n]]

Signal Amplitude Signal Amplitude

x[n] A
A x[n]
A x[−n]
A A −x[−n]
xev [n] A 2 2
2
xod [n]

A A A A n
2 2 A 2 2
2 −3 −2 −1 1 2 3

− A2
−A
2
−A
2

n −A
−3 −2 −1 1 2 3

(a) Even component xev [n] (b) Odd component xod [n]

Figure 8: Stem plots of the resulting even (left, symmetric) and odd (right, anti-symmetric) components
of x[n]. Note that xev [n] + xod [n] = x[n].

5
 Part 2: Energy and Power
2
X
E= |x[n]|2 = 3A2
n=0

3A2
Pav = lim =0
N →∞ 2N + 1

Answer: E = 3A2 , Pav = 0 (energy signal)

Example: [ 5. Properties of the Continuous-Time Unit Impulse Function] Problem: The problem
consists of two parts that investigate key properties of the Dirac delta function.

1. Evaluate the following integral: Z ∞

x(t) = δ(t − σ) dσ
0

2. Show that the time-scaled unit impulse function δ(2t) is equivalent to:
1
δ(2t) = δ(t)
2

Solution:
Part 1: Let τ = t − σ, then dτ = −dσ:
Z ∞ Z t
x(t) = δ(t − σ) dσ = δ(τ ) dτ = u(t)
0 −∞

(
1, t≥0
Continuous-Time Unit Step Function u(t) =
0, t<0

u(t)

t
−2 −1 1 2 3

Figure 9: The continuous-time unit step function, u(t). The function is 0 for t < 0 and 1 for t > 0.

Part 2: Model δ(t) as limit of rectangular pulse gT (t) with area 1. Time scaling: gT (2t) has width
T /2 and height 1/T , so area = 1/2. Therefore: δ(2t) = 12 δ(t)

6
 gT (t)

1
T

Area = 1
t
− T2 T
2

gT (2t)

1
T

1
Area = 2

− T4 T
4

Figure 10: Illustration of the rectangular pulse gT (t) and its time-scaled version gT (2t). The scaled pulse
is half as wide, and therefore has half the area.

Answer: (1) u(t), (2) δ(2t) = 21 δ(t)

7
 Lecture 3 Examples
Signals and Systems Course

Example: [ 1. Determining if a System has Memory] Problem: Determine whether the system y[n] =
(2x[n] − x[n]2 )2 is memoryless.

+
x[n] 2 (·)2 y[n]
−

(·)2

2
System Equation: y[n] = 2x[n] − (x[n])2

Figure 1: Block diagram representation of the system.

Solution:
The output y[n] depends only on the current input x[n] at time n. There is no dependence on past
or future values.
Answer: The system is memoryless.

Example: [ 2. System Properties - Memory and Invertibility] Problem: Determine which systems have
memory and which are invertible:

1. y(t) = x(t) + 1 (continuous-time)

2. y[n] = x[n − 1] (discrete-time)
3. y(t) = x2 (t)
Solution:

ˆ System 1: Memoryless (depends only on current input)

ˆ System 2: Has memory (depends on past input)

ˆ System 3: Not invertible (x1 = 2 and x2 = −2 both give y = 4)

Answer: System 1 is memoryless; System 2 has memory; System 3 is not invertible.

Example: [ 3. System Invertibility Analysis] Problem: Determine the invertibility of:

1
 1. y[n] = 2x[n] (discrete-time)
Rt
2. y(t) = −∞ x(τ )dτ (continuous-time integrator)

3. y[n] = x2 [n] (discrete-time)

Solution:
ˆ System 1: Invertible (x[n] = y[n]/2)
ˆ System 2: Invertible (x(t) = dy(t)/dt)
ˆ System 3: Not invertible (x1 = 3 and x2 = −3 both give y = 9)
Answer: Systems 1 and 2 are invertible; System 3 is not invertible.

Example: [ 4. System Causality Analysis] Problem: Determine the causality of:

1. y[n] = x[n] + x[n − 1] (discrete-time)

2. y[n] = x[n + 1] (discrete-time)
3. y(t) = x(t − 2) (continuous-time)
4. y(t) = x(t + 1) (continuous-time)
Solution:
ˆ System 1: Causal (current and past inputs)
ˆ System 2: Non-causal (future input)
ˆ System 3: Causal (past input)
ˆ System 4: Non-causal (future input)
Answer: Systems 1 and 3 are causal; Systems 2 and 4 are non-causal.

Example: [ 5. Testing for Time-Invariance] Problem: Determine if y(t) = sin[x(t)] is time-invariant.

Solution: Test: If x1 (t) → y1 (t) = sin[x1 (t)], then x1 (t − t0 ) → y2 (t) = sin[x1 (t − t0 )] Compare:
y1 (t − t0 ) = sin[x1 (t − t0 )] = y2 (t)
Answer: Time-invariant (sine operates on instantaneous values)

Example: [ 6. Testing a System for Time-Invariance] Problem: Determine if y[n] = nx[n] is time-
invariant.
Solution: Test: If x1 [n] → y1 [n] = nx1 [n], then x1 [n − n0 ] → y2 [n] = nx1 [n − n0 ] Compare:
y1 [n − n0 ] = (n − n0 )x1 [n − n0 ] ̸= y2 [n] = nx1 [n − n0 ]
Answer: Time-varying (scaling factor depends on time index n)

Example: [ 7. Testing a Time-Scaling System for Time-Invariance] Problem: Determine if y[n] = x[2n]
(decimator) is time-invariant.
Solution: Test: If x1 [n] → y1 [n] = x1 [2n], then x1 [n − n0 ] → y2 [n] = x1 [2n − n0 ] Compare:
y1 [n − n0 ] = x1 [2n − 2n0 ] ̸= y2 [n] = x1 [2n − n0 ]
Answer: Time-varying (decimation samples at specific indices that change with shift)

2
 P∞
Example: [ 8. Proving Linearity for an Ideal Sampling System] Problem: Determine if y(t) = n=−∞ x(t)δ(t−
nT ) is linear. P∞ P∞
Solution: Homogeneity: T {αx(t)} P∞ = n=−∞ αx(t)δ(t − nT ) = α n=−∞ x(t)δ(t − nT ) = αy(t)
Additivity: T {x1 (t) + x2 (t)} = n=−∞ [x1 (t) + x2 (t)]δ(t − nT ) = y1 (t) + y2 (t)
Answer: Linear (satisfies both homogeneity and additivity)

Example: [ 9. Testing a System for Linearity] Problem: Determine if y[n] = |x[n]| is linear.
Solution: Additivity: |x1 [n] + x2 [n]| ̸= |x1 [n]| + |x2 [n]| (e.g., |1 + (−1)| = 0 ̸= 2 = |1| + | − 1|)
Homogeneity: |αx[n]| = |α| · |x[n]| ̸= α|x[n]| when α < 0
Answer: Non-linear (fails both additivity and homogeneity)

Example: [ 10. Systems Exhibiting Partial Linearity Properties] Problem: Determine linearity of:

x[n]x[n−1]
1. y[n] = x[n+1]

2. y[n] = Re{x[n]}
Solution: System 1: Homogeneous (α cancels out) but not additive (nonlinear operations) System
2: Additive but not homogeneous (fails for complex α)
Answer: Both non-linear (need both additivity and homogeneity)

Example: [ 11. Stability Analysis of a Continuous-Time Integrator] Problem: Determine if y(t) =

Rt
−∞
x(τ ) dτ is stable.
Solution: Impulse R ∞ response: Rh(t) = u(t) (apply δ(t) as input)
∞
Stability test: −∞ |h(t)| dt = 0 1 dt = ∞
Answer: Unstable (impulse response not absolutely integrable)

3
 Lecture 4 Examples
Signals and Systems Course

Example: [ 1. Discrete-Time Convolution: Exponential and Unit Step] Problem: Find the convolution
y[n] = x[n] ∗ h[n] where:

x[n] = an u[n], |a| < 1 (1)

h[n] = u[n] (2)

Causal Decaying Exponential Sequence: x[n] = an u[n], for 0 < a < 1 Impulse Response of an Accumulator: h[n] = u[n]

x[n] h[n]
1
1
a 1 1 1 1 1 1 1
1
a2
a3
a4
a5

n n
−2 −1 1 2 3 4 5 6 −3 −2 −1 1 2 3 4 5 6

Figure 1: Input signal x[n] = an u[n] Figure 2: Impulse response h[n] = u[n]

Solution:
∞ n
X X 1 − an+1
y[n] = x[k]h[n − k] = ak =
1−a
k=−∞ k=0

For n < 0: y[n] = 0

Answer:
1 − an+1
y[n] = u[n]
1−a

1
 1−an+1
Convolution Result: y[n] = (an u[n]) ∗ u[n] = 1−a u[n]

y[n] 1
Asymptote: y → 1−a
1
1−a

a2
+
a
+
1
a
+
1

1
1

n
−2 −1 1 2 3 4 5 6 7 8 9 10

Figure 3: Result of convolution y[n] = x[n] ∗ h[n]

Example: [ 2. Convolution of Two Exponential Sequences] Problem: Find the convolution y[n] =
x[n] ∗ h[n] where:

x[n] = an u[n], |a| < 1 (3)

h[n] = bn u[n], |b| < 1 (4)

Solution:
n n  k
X X a
y[n] = ak bn−k = bn
b
k=0 k=0

Case 1: a ̸= b

n+1
n1 − ab bn+1 − an+1
y[n] = b a =
1− b b−a
Case 2: a = b
y[n] = an (n + 1)
Answer: (
bn+1 −an+1
b−a u[n] if a ̸= b
y[n] = n
a (n + 1)u[n] if a = b

Example: [ 3. Convolution of Right-Sided and Left-Sided Signals] Problem: Find the convolution
y[n] = x[n] ∗ h[n] where:

x[n] = u[n] (5)

 n
1
h[n] = u[n] (6)
2
Solution:
n  n−k n
 n X  n n+1
X 1 1 1 2 −1
y[n] = = k
2 = = 2 − 2−n
2 2 2 2−1
k=0 k=0

2
 For n < 0: y[n] = 0
Answer:
y[n] = (2 − 2−n )u[n]

Example: [ 4. Finite-Duration Convolution] Problem: Find the convolution y[n] = x[n] ∗ h[n] where:
(
1 for 0 ≤ n ≤ 2
x[n] = (7)
0 otherwise
(
n for 0 ≤ n ≤ 2
h[n] = (8)
0 otherwise

Finite-Duration Rectangular Pulse: x[n] = u[n] − u[n − 3] Finite-Duration Ramp: h[n] = n(u[n] − u[n − 3])

x[n] h[n]
2
2
1 1 1
1

1
1

n n
−2 −1 1 2 3 4 −2 −1 1 2 3 4

Figure 4: Input signal x[n] Figure 5: Impulse response h[n]

Solution:
For each n, sum over the overlap region max(0, n − 2) ≤ k ≤ min(2, n):

y[0] = 0 (9)
y[1] = 1 · 1 = 1 (10)
y[2] = 1 · 2 + 1 · 1 + 1 · 0 = 3 (11)
y[3] = 1 · 2 + 1 · 1 = 3 (12)
y[4] = 1 · 2 = 2 (13)

Answer: 

0 for n < 0 or n > 4

1 for n=1
y[n] =


3 for n = 2, 3
2 for n=4


3
 Convolution result y[n] = x[n] ∗ h[n]

3 y[n]

n
0
0 1 2 3 4 5

Figure 6: Result of convolution y[n] = x[n] ∗ h[n]

Example: [ 5. Convolution of Two Finite-Length Sequences] Problem: Find y[n] = x[n] ∗ h[n] where:

1
 n=0
h[n] = −1/2 n=1

0 otherwise




1 n = −1

2 n=0
x[n] =


−1 n=1
0 otherwise


Discrete-Time Input Signal x[k]

x[k]
2

1
1

k
−2 −1 1 2 3

−1
−1

Figure 7: The finite-length input signal x[k].

Solution: y[n] is non-zero for n ∈ {−1, 0, 1, 2}:

y[−1] = x[−1]h[0] = (1)(1) = 1 (14)

y[0] = x[−1]h[1] + x[0]h[0] = (1)(−1/2) + (2)(1) = 3/2 (15)
y[1] = x[0]h[1] + x[1]h[0] = (2)(−1/2) + (−1)(1) = −2 (16)
y[2] = x[1]h[1] = (−1)(−1/2) = 1/2 (17)

4
 Alternative: Decompose x[n] = δ[n + 1] + 2δ[n] − δ[n − 1] Then y[n] = h[n + 1] + 2h[n] − h[n − 1]
gives same result.

Discrete-Time Output Signal y[n]

y[n]
1.5
1.5
1
1
0.5
n
−2 −1 1 2 3

−2
−2

Figure 8: The final output signal y[n].

Answer: y[n] = {1, 3/2, −2, 1/2} for n = {−1, 0, 1, 2}, 0 otherwise

5
 Lecture 5 Examples
Signals and Systems Course

Example: [ 1. Continuous-Time Convolution: Exponential and Unit Step] Problem: Find the convo-
lution y(t) = x(t) ∗ h(t) where:
x(t) = u(t) (unit step function) (1)
−at
h(t) = e u(t), a>0 (exponential impulse response) (2)

Impulse Response h(t) = e−at u(t), for a > 0

(
1, t≥0
Unit Step Function x(t) = u(t) =
0, t<0
h(t)
x(t) 1

e−1

t
t
1
−1 1 2 3 4 a

Figure 1: Input signal x(t) = u(t) Figure 2: Impulse response h(t) = e−at u(t)

Solution:
t t
1 − e−at
Z Z
y(t) = e−a(t−τ ) dτ = e−at eaτ dτ =
0 0 a
For t < 0: y(t) = 0
Answer:
1 − e−at
y(t) = u(t)
a

Step Response: y(t) = u(t) ∗ e−at u(t) = (1 − e−at )u(t)

y(t)
1

1 − e−1

t
1
a

Figure 3: Step response y(t) = x(t) ∗ h(t)

1
Example: [ 2. Convolution of Two-Sided and One-Sided Signals] Problem: Find the convolution y(t) =
x(t) ∗ h(t) where:
x(t) = e−|t| (two-sided exponential) (3)
h(t) = u(t) (unit step function) (4)

Two-Sided Exponential Signal: x(t) = e−a|t| Impulse Response h(t) = u(t)

x(t) h(t)
1

e−1

t t
− a1 1
a −2 −1 1 2 3

Figure 4: Input signal x(t) = e−|t| Figure 5: Impulse response h(t) = u(t)

Solution:
Z t
y(t) = e−|τ | dτ
−∞
Case 1: t < 0 Z t
y(t) = eτ dτ = et
−∞
Case 2: t ≥ 0 Z 0 Z t
y(t) = eτ dτ + e−τ dτ = 1 + (1 − e−t ) = 2 − e−t
−∞ 0
Answer: (
et for t < 0
y(t) =
2 − e−t for t ≥ 0

(
et , t<0
Convolution Result y(t) =
2 − e−t , t ≥ 0

y(t)

2 − e−t

et

t
−2 −1 1 2

Figure 6: Convolution result y(t) = x(t) ∗ h(t)

2
Example: [ 3. Convolution with Rectangular Pulse] Problem: Find the convolution y(t) = x(t) ∗ h(t)
where:

x(t) = e−at u(t), a > 0 (5)

(
1 for 0 ≤ t ≤ T
h(t) = (6)
0 otherwise

Impulse Response h(t) = A[u(t) − u(t − T )]

Input Signal x(t) = Ae−at u(t), for a > 0
h(t)
x(t)
A
A

Ae−1

t t
1 T
a

Figure 8: Impulse response h(t) (rectangular

Figure 7: Input signal x(t) = e−at u(t)
pulse)

Solution:
Z t
y(t) = e−aτ dτ
max(0,t−T )

Case 1: t < 0 
y(t) = 0
Case 2: 0 ≤ t ≤ T
t
1 − e−at
Z
y(t) = e−aτ dτ =
0 a
Case 3: t > T
t
e−a(t−T ) − e−at e−at (eaT − 1)
Z
y(t) = e−aτ dτ = =
t−T a a
Answer: 
0
 for t < 0
1−e−at
y(t) = a for 0 ≤ t ≤ T
 e−at (eaT −1)

a for t > T

3
 Convolution: (e−at u(t)) ∗ [u(t) − u(t − T )]

y(t)

1 − e−aT

1 − e−at e−at (eaT − 1)

for 0 ≤ t ≤ T for t > T

t
T

Figure 9: Convolution result y(t) = x(t) ∗ h(t)

Example: [ 4. Convolution with Impulse Train] Problem: Find y(t) = x(t) ∗ h(t) where x(t) = cos(ω0 t)
P∞
and h(t) = n=−∞ δ(t − nT ). P∞
Solution: Using sifting property: y(t) = n=−∞ cos(ω0 (t − nT ))
Using the identity cos(A − B) = cos A cos B + sin A sin B:
∞
X
y(t) = [cos(ω0 t) cos(ω0 nT ) + sin(ω0 t) sin(ω0 nT )]
n=−∞

∞
X ∞
X
= cos(ω0 t) cos(ω0 nT ) + sin(ω0 t) sin(ω0 nT )
n=−∞ n=−∞
P∞
Since n=−∞ sin(ω0 nT ) = 0 (odd symmetry):
∞
X
y(t) = cos(ω0 t) cos(ω0 nT )
n=−∞
P∞
Answer: y(t) = cos(ω0 t) n=−∞ cos(ω0 nT )

4
 Lecture 6 Examples
Signals and Systems Course

Example: [ 1. Finding an LTI System’s Output Using its Impulse Response] Problem: Given LTI
system responses:
ˆ x [n] = δ[n] + δ[n − 1] → y [n] =
1 1
1
2 (δ[n] + δ[n − 1] − δ[n − 2] − δ[n − 3])
ˆ x [n] = δ[n] − δ[n − 1] → y [n] =
2 2
1
2 (δ[n] − δ[n − 1] + δ[n − 2] − δ[n − 3])
Find the output for x[n] = cos(πn).

Input-Output Pair 1

x1 [n] Amplitude
y1 [n] 1

0.5
n
−1 1 2 3 4
−0.5

−1

Input-Output Pair 2

x2 [n] Amplitude
y2 [n] 1

0.5
n
−1 1 2 3 4
−0.5

−1

Figure 1: The two known input-output pairs for the LTI system S.

Solution:
Step 1: Find impulse response
1
δ[n] = (x1 [n] + x2 [n])
2
1 1
h[n] = (y1 [n] + y2 [n]) = (δ[n] − δ[n − 3])
2 2

1
 Step 2: Find output for cos(πn) = (−1)n
1
y[n] = h[n] ∗ (−1)n = (δ[n] − δ[n − 3]) ∗ (−1)n
2
∞
1 X
= (−1)n (δ[k] − δ[k − 3])(−1)−k
2
k=−∞

1
= (−1)n [1 − (−1)] = (−1)n
2
Answer: y[n] = (−1)n = cos(πn)

Example: [ 2. Inverse Systems and System Cascading] Problem: Given LTI systems:

ˆ System S : y [n] = x[n − 1] (delay of 1)

1 1

ˆ System S : y [n] = x[n + 1] (advance of 1)

2 2

Find the overall system response when cascaded.

Solution:
Impulse responses:

h1 [n] = δ[n − 1] (1)

h2 [n] = δ[n + 1] (2)

Overall impulse response:

htotal [n] = h1 [n] ∗ h2 [n] = δ[n − 1] ∗ δ[n + 1] = δ[n]

Answer: y[n] = x[n] (identity system)

Example: [ 3. First-Difference Filter as Inverse of Accumulator] Problem: Show that the first-difference
filter is the inverse of the accumulator:
ˆ Pn
Accumulator: y[n] = k=−∞ x[k]
ˆ First-difference filter: y[n] = x[n] − x[n − 1]
Solution:
Impulse responses:

hacc [n] = u[n] (3)

hdif f [n] = δ[n] − δ[n − 1] (4)

Cascaded system:

htotal [n] = u[n] ∗ (δ[n] − δ[n − 1]) = u[n] − u[n − 1] = δ[n]

Answer: The first-difference filter is the inverse of the accumulator.

Example: [ 4. Causality of LTI Systems] Problem: Determine which systems are causal:

1. h1 (t) = e−t u(t)

2. h2 (t) = e−|t|

2
 3. h3 (t) = δ(t + 1)
4. h4 (t) = e−t u(−t)
Solution:
An LTI system is causal if h(t) = 0 for all t < 0.

ˆ System 1: h (t) = 0 for t < 0  Causal

1

ˆ System 2: h (t) = e ̸= 0 for t < 0  Non-causal

2
t

ˆ System 3: h (t) ̸= 0 at t = −1 < 0  Non-causal

3

ˆ System 4: h (t) = e ̸= 0 for t < 0  Non-causal

4
−t

Answer: Only System 1 is causal.

Example: [ 5. Finding Impulse Response from Integral Equation] Problem: Find the impulse response
of the LTI system: Z t
y(t) = e−(t−τ ) x(τ )dτ
−∞

Solution:
Z t
h(t) = e−(t−τ ) δ(τ )dτ
−∞

For t < 0: h(t) = 0 (no overlap with δ(τ )) For t ≥ 0: h(t) = e−t
Answer: h(t) = e−t u(t)

Example: [ 6. BIBO Stability of LTI Systems] Problem: Determine BIBO stability of:

1. h1 [n] = an u[n], |a| < 1

2. h2 [n] = an u[n], |a| ≥ 1

3. h3 (t) = e−at u(t), a > 0
4. h4 (t) = eat u(t), a > 0


Solution: P
∞ 1
System 1: |a|n = 1−|a| <∞ Stable
Pn=0
∞ n
System 2: R n=0 |a| = ∞ 

Not stable
∞
System 3: 0 e−at dt = a1 < ∞ Stable
R∞
System 4: 0 eat dt = ∞ 
Not stable
Answer: Systems 1 and 3 are BIBO stable; Systems 2 and 4 are not.

Example: [ 7. Autocorrelation of a One-Sided Exponential Signal] Problem: Find Rxx (t) for x(t) =
R∞
2e−3t u(t) where Rxx (t) = −∞ x(τ )x(τ − t)dτ .

3
 Exponential Signal x(τ ) = Ae−aτ u(τ )

x(τ )
A

Ae−1

τ
1
a

Figure 2: The signal x(τ ) = 2e−3τ u(τ ).

Solution: Z ∞
Rxx (t) = 4e3t
e−6τ u(τ )u(τ − t)dτ
−∞
Case 1: t ≥ 0 (overlap when τ ≥ t):
Z ∞
2 −3t
Rxx (t) = 4e 3t
e−6τ dτ = e
t 3
Case 2: t < 0 (overlap when τ ≥ 0):
Z ∞
2 3t
Rxx (t) = 4e3t e−6τ dτ = e
0 3

Autocorrelation Function Rxx (t) = 23 e−3|t|

Rxx (t)
2
3

t
−1 1

Figure 3: The resulting autocorrelation function is a symmetric, two-sided exponential, which is maximal
at t = 0.

Answer: Rxx (t) = 23 e−3|t|

Example: [ 8. Cross-Correlation of Two Rectangular Pulses] Problem: Given two continuous-time

rectangular pulse signals, x(t) and y(t), defined as:

4
 ˆ x(t) = u(t) − u(t − 1)
ˆ y(t) = u(t − 3/2) − u(t − 5/2)
Find the cross-correlation functions Ryx (t) and Rxy (t). These are defined by the integrals:
Z ∞
Ryx (t) = y(τ )x(τ − t)dτ
−∞
Z ∞
Rxy (t) = x(τ )y(τ − t)dτ
−∞

Rectangular Pulse Signals

Amplitude
x(t)
y(t)

t
1 1.5 2.5

Figure 4: The two rectangular pulse signals x(t) and y(t).

Solution: The cross-correlation integral represents the area of overlap between one signal and a
time-shifted version of the other. Since both signals are pulses of height 1, the value of the integral is
simply the length of the interval over which they overlap.
Calculation of Rxy (t): We need to find the overlap between x(τ ) (which exists for 0 ≤ τ ≤ 1) and
y(τ − t) (which exists for 3/2 ≤ τ − t ≤ 5/2, or 3/2 + t ≤ τ ≤ 5/2 + t).
The overlap occurs when both conditions are satisfied:
ˆ 0 ≤ τ ≤ 1 (from x(τ ))
ˆ 3/2 + t ≤ τ ≤ 5/2 + t (from y(τ − t))
This gives us different cases based on the value of t:
Case 1: t ≤ −5/2 - No overlap, Rxy (t) = 0
Case 2: −5/2 < t ≤ −3/2 - Partial overlap The overlap interval is [0, 5/2 + t], so:

Rxy (t) = 5/2 + t

Case 3: −3/2 < t ≤ −1/2 - Partial overlap The overlap interval is [3/2 + t, 1], so:

Rxy (t) = 1 − (3/2 + t) = −1/2 − t

Case 4: t ≥ −1/2 - No overlap, Rxy (t) = 0

5
 Cross-Correlation Function Rxy (t)

Rxy (t)
1

t
−1.5 −0.5 0.5 1.5

Figure 5: The two cross-correlation functions.

Answer: The cross-correlation functions are:



 0 t ≤ −5/2

5/2 + t −5/2 < t ≤ −3/2
Rxy (t) =


 −1/2 − t −3/2 < t < −1/2
0 t ≥ −1/2


And by the property of cross-correlation: Ryx (t) = Rxy (−t)

Example: [ 9. Autocorrelation of a Discrete-Time Exponential Sequence] Problem: For the discrete-

time signal x[n], defined as a one-sided exponential sequence:
 n
1
x[n] = u[n]
2
Find its autocorrelation function, Rxx [n], which is defined by the summation:
∞
X
Rxx [n] = x[k]x[k − n]
k=−∞

Discrete Exponential Signal x[k] = ak u[k], for a = 1/2

x[k]
1
1

a2

a3
a4
k
−2 −1 1 2 3 4 5 6

Figure 6: The signal x[k] = (1/2)k u[k].

6
 Solution: We substitute the definition of x[n] into the autocorrelation summation:
∞  k  k−n
X 1 1
Rxx [n] = u[k] u[k − n]
2 2
k=−∞

Simplifying:
∞  2k
 −n X
1 1
Rxx [n] = u[k]u[k − n]
2 2
k=−∞

The product u[k]u[k − n] is non-zero only when both step functions are 1, which occurs when k ≥ 0
and k ≥ n. This gives us two cases:
Case 1: n ≥ 0 For n ≥ 0, both conditions are satisfied when k ≥ n. The summation becomes:
 −n X ∞  2k  −n  2n X ∞  k
1 1 1 1 1
Rxx [n] = =
2 2 2 2 4
k=n k=0

Using the geometric series formula:

 n  n  n
1 1 1 4 4 1
Rxx [n] = · = · =
2 1 − 1/4 2 3 3 2
Case 2: n < 0 For n < 0, both conditions are satisfied when k ≥ 0. The summation becomes:
 −n X ∞  2k  −n  −n
1 1 1 1 4 1
Rxx [n] = = · =
2 2 2 1 − 1/4 3 2
k=0

Discrete Autocorrelation Rxx [n] = 34 ( 12 )|n|

4
Rxx [n]
1.33

0.67 0.67

0.33 0.33

0.17 0.17

n
−3 −2 −1 1 2 3

Figure 7: The resulting autocorrelation function is a symmetric, two-sided exponential sequence.

Answer: The autocorrelation function is:

4 1 n
(
 |n|
3 2
n≥0 4 1
Rxx [n] = 4 1 −n =
3 2 n<0 3 2

Example: [ 10. Cross-Correlation of Two Discrete-Time Signals] Problem: Compute the cross-correlation
Rxy [n] for the two discrete-time signals x[n] and y[n]:
The signals can be expressed mathematically as:
x[n] = 2δ[n + 1] + 4δ[n] − 3δ[n − 1]
y[n] = −3δ[n + 1] + δ[n − 1]
where δ[n] is the unit impulse function. The non-zero values are:

7
 ˆ x[−1] = 2, x[0] = 4, x[1] = −3
ˆ y[−1] = −3, y[0] = 0, y[1] = 1
Discrete-Time Signals
4 Amplitude
x[n] 4
y[n]
2
2
1

n
−2 −1 1 2

−3 −3 −3

Figure 8: The two discrete-time signals x[n] and y[n].

Solution: The cross-correlation is defined as:

∞
X
Rxy [n] = x[k]y[k − n]
k=−∞

Since both signals have finite support, we only need to consider the non-zero terms. Let’s compute
Rxy [n] for different values of n:
For n = −2:
X
Rxy [−2] = x[k]y[k + 2] = x[−1]y[1] + x[0]y[2] + x[1]y[3] = 2 · 1 + 4 · 0 + (−3) · 0 = 2
k

For n = −1:
X
Rxy [−1] = x[k]y[k + 1] = x[−1]y[0] + x[0]y[1] + x[1]y[2] = 2 · 0 + 4 · 1 + (−3) · 0 = 4
k

For n = 0:
X
Rxy [0] = x[k]y[k] = x[−1]y[−1] + x[0]y[0] + x[1]y[1] = 2 · (−3) + 4 · 0 + (−3) · 1 = −9
k

For n = 1:
X
Rxy [1] = x[k]y[k − 1] = x[−1]y[−2] + x[0]y[−1] + x[1]y[0] = 2 · 0 + 4 · (−3) + (−3) · 0 = −12
k

For n = 2:
X
Rxy [2] = x[k]y[k − 2] = x[−1]y[−3] + x[0]y[−2] + x[1]y[−1] = 2 · 0 + 4 · 0 + (−3) · (−3) = 9
k

For all other values of n, Rxy [n] = 0.

8
 Discrete Cross-Correlation Rxy [n]
9
9 Rxy [n]

4
2 4
2
n
−3 −2 −1 1 2 3

−9
−9
−12
−12

Figure 9: The cross-correlation function Rxy [n].

Answer: The cross-correlation function is:



 2 n = −2

4 n = −1





−9 n=0
Rxy [n] =


 −12 n=1
9 n=2




0 otherwise


Example: [ 11. Autocorrelation of a Continuous-Time Rectangular Pulse] Problem: Given the continuous-
time rectangular pulse signal:
x(t) = u(t) − u(t − 1)
where u(t) is the continuous-time unit step function.
Find the autocorrelation function of this signal, denoted by Rxx (t), which is defined by the integral:
Z ∞
Rxx (t) = x(τ )x(τ − t) dτ
−∞

Rectangular Pulse x(t) = u(t) − u(t − T )

x(t)

t
T

Figure 10: The rectangular pulse signal x(t) = u(t) − u(t − 1).

9
 Solution: The autocorrelation function Rxx (t) can be interpreted as the convolution of x(t) with
its time-reversed version, x(−t). For a real-valued signal like this one, it measures the area of overlap
between the signal x(τ ) and a version of itself, x(τ − t), that has been shifted by an amount t.
Intuitive Approach: Graphical ”Slide and Overlap” We can visualize the calculation by fixing
the signal x(τ ) in place and ”sliding” a shifted copy, x(τ − t), across it. The value of the autocorrelation
Rxx (t) for any given shift t is simply the area of the overlapping region between the two pulses.
ˆ The fixed pulse x(τ ) occupies the interval [0, 1].
ˆ The sliding pulse x(τ − t) occupies the interval [t, 1 + t].
ˆ The autocorrelation R (t) will be the area of the intersection of these two intervals.
xx

This graphical method shows that the overlap will only occur for shifts between t = −1 and t = 1.
Outside this range, the pulses do not overlap, and the autocorrelation is zero. The overlap area will be
maximum (equal to 1) at t = 0 and will decrease linearly to zero as t approaches −1 or 1. This predicts
a triangular shape for the result.
Mathematical Calculation We need to evaluate the integral:
Z ∞
Rxx (t) = x(τ )x(τ − t) dτ
−∞

Since x(τ ) = 1 for 0 ≤ τ ≤ 1 and 0 otherwise, and x(τ − t) = 1 for t ≤ τ ≤ 1 + t and 0 otherwise, the
integrand is non-zero only when both conditions are satisfied.
Case 1: t ≤ −1 No overlap occurs, so Rxx (t) = 0.
Case 2: −1 < t ≤ 0 The overlap interval is [0, 1 + t], so:
Z 1+t
Rxx (t) = 1 · 1 dτ = 1 + t
0

Case 3: 0 < t ≤ 1 The overlap interval is [t, 1], so:

Z 1
Rxx (t) = 1 · 1 dτ = 1 − t
t

Case 4: t > 1 No overlap occurs, so Rxx (t) = 0.

Autocorrelation of a Rectangular Pulse Rxx (t) = (1 − |t|)rect( 2t )

Rxx (t)
1

t
−1 1

Figure 11: The resulting triangular autocorrelation function Rxx (t).

Answer: The autocorrelation function is:



 0 t ≤ −1

1 + t −1 < t ≤ 0
Rxx (t) = = max(0, 1 − |t|)


 1−t 0<t≤1
0 t>1


10
Example: [ 12. Fundamental Properties of Correlation Functions] Problem: This problem consists of
two independent proofs regarding correlation functions.

1. Autocorrelation Maximum Property: Show that for any real-valued energy signal x(t), its
autocorrelation function, Rxx (t), satisfies the following inequality for all t:

Rxx (0) ≥ Rxx (t)

2. Correlation-Convolution Relationship: Show that the cross-correlation of two real-valued

signals, x(t) and y(t), can be expressed as the convolution of one signal with the time-reversed
version of the other:
Rxy (t) = x(t) ∗ y(−t)

Solution: We will prove each statement step-by-step.

Proof 1: Autocorrelation Maximum at the Origin This property states that a signal is always
most similar to itself when there is no time shift. R∞
Intuitive Approach: The autocorrelation function, Rxx (t) = −∞ x(τ )x(τ − t) dτ , measures the
area
R ∞ of2 the product of a signal and a time-shifted version of itself. At t = 0, the function is Rxx (0) =
2
−∞
x (τ ) dτ , which is the total energy of the signal. Since x (τ ) is always non-negative, this integral
represents the maximum possible overlap area. Any shift (t ̸= 0) is likely to reduce the alignment between
positive and negative parts of the signal, thus reducing the value of the integral.
Mathematical Proof: Consider the integral:
Z ∞
[x(τ ) − x(τ − t)]2 dτ ≥ 0
−∞

Expanding the square:

Z ∞
[x2 (τ ) − 2x(τ )x(τ − t) + x2 (τ − t)] dτ ≥ 0
−∞

Using the linearity of integration, we can split this into three separate integrals:
Z ∞ Z ∞ Z ∞
x2 (τ ) dτ − 2 x(τ )x(τ − t) dτ + x2 (τ − t) dτ ≥ 0
−∞ −∞ −∞

We now identify each of these terms based on the definition of autocorrelation:

ˆR∞ 2
−∞
x (τ ) dτ = Rxx (0) (This is the autocorrelation at zero lag, i.e., the signal’s energy).

ˆR∞
−∞
x(τ )x(τ − t) dτ = Rxx (t) (This is the definition of the autocorrelation function).

ˆR∞ 2
−∞
x (τ − t) dτ = Rxx (0) (The energy of a signal is invariant to a time shift).
Substituting these into our inequality:

Rxx (0) − 2Rxx (t) + Rxx (0) ≥ 0

2Rxx (0) − 2Rxx (t) ≥ 0

Rxx (0) − Rxx (t) ≥ 0
Therefore:
Rxx (0) ≥ Rxx (t)

Proof 2: Relationship Between Correlation and Convolution We need to show that Rxy (t) =
x(t) ∗ y(−t).
Starting with the definition of cross-correlation:
Z ∞
Rxy (t) = x(τ )y(τ − t) dτ
−∞

11
 Let’s make a substitution: let u = τ − t, so τ = u + t and dτ = du. When τ → −∞, u → −∞, and
when τ → ∞, u → ∞.
Z ∞
Rxy (t) = x(u + t)y(u) du
−∞

Now let’s make another substitution: let v = −u, so u = −v and du = −dv. When u → −∞, v → ∞,
and when u → ∞, v → −∞.
Z −∞ Z ∞
Rxy (t) = x(−v + t)y(−v)(−dv) = x(t − v)y(−v) dv
∞ −∞

This is exactly the definition of convolution between x(t) and y(−t):

Rxy (t) = x(t) ∗ y(−t)

Therefore:
Rxy (t) = x(t) ∗ y(−t)

Answer: Both properties have been proven:

1. The autocorrelation function reaches its maximum at zero lag: Rxx (0) ≥ Rxx (t) for all t.
2. The cross-correlation can be expressed as a convolution: Rxy (t) = x(t) ∗ y(−t).

12