EECE 340 Signals and Systems Feb.

5, 2024
American University of Beirut, Spring 2023–2024 Handout #4

Problem Set 1 – Solutions

Problem 1.1

1) Determining the difference equation from the diagram, we have

y[n] = α(x[n − 1] − y[n − 2]).

Since X is an impulse and the system is at rest, from the difference equation we
note that
y[0] = 0, y[1] = α, y[2] = 0, y[3] = −α2 , · · ·

Therefore, we conclude that α = 0.5.

2) The system does not have a finite impulse response based on what we identified
earlier. Indeed, we claim that
(−1)(n−1)/2 α(n+1)/2 n is positive and odd

y[n] =
0 o.w.

We prove the claim by induction:

(a) It is true for n = 0, 1.
(b) Assume it is true up to n − 1 ≥ 1 and we prove that this implies that it is true
for n. If n is even then both x[n − 1] and y[n − 2] are zero. If n is odd, then
x[n − 1] = 0 and

y[n] = −αy[n − 2] = −(−1)(n−3)/2 αα(n−1)/2 = (−1)(n−1)/2 α(n+1)/2 .

Finally, since α is positive, then y[n] is non-zero for arbitrarily large indices n and
y[n] is not finite length.

Problem 1.2

(a) Using the proposed relationship

ay[n] + by[n − 1] = cx[n] + dx[n − 1],

we note that for any solution (a, b, c, d), say (2a, 2b, 2c, 2d) is also a solution. Said
differently, the solution is really a “space” where all numbers can be scaled by an
arbitrary constant.

1
 Keeping this in mind, we will find the solution for which a = 1, i.e., one of the form:

y[n] + by[n − 1] = cx[n] + dx[n − 1].

Now identifying with the given functions:

n = 0; 0=c
n = 1; 0.1 = c + d
n = 2; 0.19 + 0.1b = c + d

and hence,
c = 0; d = 0.1; b = −0.9,
yielding
y[n] = 0.9y[n − 1] + 0.1x[n − 1].

(b) A block diagram of the system is readily drawn:

0.9 ✛

X ❄
✲Y
✲ 0.1 ✲ + ✲ Delay

Problem 1.3

To solve this problem we will rely on the fact that the system at hand is linear : The
output y[n] is due to the input x[n] in “addition” to the behavior due to the initial
conditions.

ˆ Initial conditions: without an input and with only the initial conditions the recur-
rence implies that 

 0 n < −2
1 n = −2




0 n = −1




1
n=0


 4

yI [n] = 0 n=1
1

 42
n=2
..






 .


 0 n odd
 1

n = 2k i.e., even
4k+1

2
 This can be readily established by induction. Indeed, it is try for n = 0 and n = 1
and assuming it is true upto n = 2k, then for n = 2k + 1 the recurrence shows that
yI [n] is zero and for n = 2(k + 1) it is (1/4) of yI [2k] which is exactly the proposed
solution.
ˆ Input: Without initial conditions, we follow the methodology developed in class and
“break” the system into fundamental ones: The system function is
Yo 1
= .
X 1 − (1/4)D2

Performing a partial fraction expansion:

Yo (1/2) (1/2)
= + .
X 1 − (1/2)D 1 + (1/2)D

We have two first-order systems in parallel. We check the response of each for the
unit-step input: For
Y 1
= , y[n] = py[n − 1] + x[n]
X 1 − pD
a unit step input will result in


 0 n<0



 1 n=0
 1+α n=1



[n] = 1 + α + α2 n=2
1−α4

 1 + α + α2 + α3 = 1−α
n=3
..





 .
 1−αk+1

1−α
n=k
which readily established by induction.
Therefore,
1 1 − (1/2)n+1 1 − (−1/2)n+1
 
yo [n] = + u[n]
2 1 − (1/2) 1 − (−1/2)
1
= 1 − (1/2)n+1 u[n] + 1 − (−1/2)n+1 u[n]



3
In summary:
  2k+1
1 4 2 1
+ − n = 2k, k ≥ 0


4k+1

 3 3 2#
y[n] = yI [n] + yo [n] = "  2k+2
4 1
1− n = 2k + 1, k ≥ 0



 3 2
  n
4 1 1
 − n even


3 12 2
= n
4 2 1
 − n odd.


3 3 2

3
Problem 1.4

1) The unit-sample or impulse defined in class is an even signal, indeed, x[−n] = x[n]
for all integers.
2) Let x[n] be a signal that is both even and odd. Examining the value at zero,
x[0] = −x[0] because it is odd and hence x[0] = 0. Actually, any odd signal has to
be equal to zero at time 0.
Additionally, because it is even x[−n] = x[n] for any n 6= 0. Being also odd means
that x[−n] = −x[n] for any such n. Therefore, x[n] = −x[n] = 0.
In conclusion, the only signal that is both even and odd is the all zero signal.
3) It is possible to have a signal that is neither such as
x[0] = 1, x[1] = 1, x[n] = 0 for all other values of n.

This signal is clearly not odd because it is not zero at time n = 0 and it is not even
because x[−1] = 0 6= x[1] = 1.
4) If we define
1
xe [n] = [x[n] + x[−n]] ,
2
then xe [n] is even:
1
xe [−n] = [x[−n] + x[n]] = xe [n],
2
since addition is commutative.
6) If we were to decompose every signal into an even and an odd part,
x[n] = xe [n] + xo [n],
1
by computing the difference xo [n] = 2
[x[−n] − x[n]], which indeed an odd signal
as xo [−n] = −xo [n].
5) This decomposition into a sum of even and odd signals is unique. Indeed, one can
make the following observations:
(a) Let us decompose the “all-zero” signal 0[n] as
0[n] = e[n] + o[n].
The odd signal o[n] is zero at n = 0 by what was derived above. Assume that
this odd signal is not the all zero signal, i.e., o[l] = c 6= 0 for some l 6= 0.
Since the sum is zero, e[l] = −c. Now observe the value of these signals at
n = −l,
0 = 0[−l] = e[−l] + o[−l] = e[l] − o[−l] = −c − c = −2c,
which implies that c = 0 which is a contradiction.
In conclusion, 0[n] can be decomposed uniquely into
0[n] = 0[n] + 0[n].

4
 (b) Now let x[n] be any signal and assume that it can be decomposed into an even
and odd function in two different ways

x[n] = e1 [n] + o1 [n]; x[n] = e2 [n] + o2 [n].

Taking the difference,

0[n] = (e1 − e2 )[n] + (o1 − o2 )[n].

Now since the linear combination of two even/odd signals is even/odd respec-
tively, and since the decomposition of 0[n] is unique, then necessarily

e1 = e2 & o1 = o2 .

7) The even part of x[n] is

 n  −n !  
1 1 1 1 1
e[n] = (x[n] + x[−n]) = + = cosh n ln .
2 2 3 3 3

Its odd part is

 n  −n !  
1 1 1 1 1
o[n] = (x[n] − x[−n]) = − = sinh n ln .
2 2 3 3 3

8) A schematic diagram is shown below

400

300

200

100

−100

−200

−300

−400
−6 −4 −2 0 2 4 6