Chapter 1 Solutions

R2 R3
1-1
1.1-1. (a) E = 0
(1)2 dt + 2 (−1)2 dt =3
R2 R3
(b) E = 0
(−1)2 dt + 2 (1)2 dt =3
R2 2 R3
(c) E = 0
(2) dt + 2 (−2)2 dt = 12
R5 2 R6
(d) E = 3
(1) dt + 5 (−1)2 dt =3
Comments: Changing the sign of a signal does not change its energy. Doubling a
signal quadruples its energy. Shifting a signal does not change its energy. Multiplying
a signal by a constant K increases its energy by a factor K 2 .
1-2
1.1-2.
Z 1 Z 0
1 3 ¯¯1 1 1 3 ¯¯0 1
Ex = t2 dt = t 0= , Ex1 = (−t)2 dt = t −1 = ,
0 3 3 −1 3 3
Z 1 2 Z Z 1
1 ¯1 1 1
Ex2 = (−t) dt = t3 ¯0 = ,
2
Ex3 = (t − 1)2 dt = x2 dx = ,
0 3 3 1 0 3
Z 1
4 ¯0 4
Ex4 = (2t)2 dt = t3 ¯−1 =
0 3 3

1.1-3. (a)
1-3
Z 2 Z 1 Z 2
2 2
Ex = (1) dt = 2, Ey = (1) dt + (−1)2 dt = 2,
0 0 1
Z 1 Z 2
2
Ex+y = (2) dt = 4, Ex−y = (2)2 dt = 4
0 1

Therefore Ex±y = Ex + Ey .
(b)
Z 2π Z 2π Z 2π
1 1
Ex = sin2 t dt = (1)dt − cos(2t)dt = π + 0 = π
0 2 0 2 0
Z 2π
Ey = (1)2 dt = 2π
0
Z 2π Z 2π Z 2π Z 2π
Ex+y = (sin t+1)2 dt = sin2 (t)dt+2 sin(t)dt+ (1)2 dt = π+0+2π = 3π
0 0 0 0

In both cases (a) and (b), Ex+y = Ex + Ey . Similarly we can show that for both
cases Ex−y = Ex + Ey .

23

1
 (c) As seen in part (a), Z π
Ex = sin2 t dt = π/2
0
Furthermore, Z π
Ey = (1)2 dt = π
0
Thus,
Z π Z π Z π Z π
2 2 3π
Ex+y = (sin t+1) dt = sin (t)dt+2 sin(t)dt+ (1)2 dt = π/2+2(2)+π +4
0 0 0 0 2

Additionally,
Z π
3π
Ex−y = (sin t − 1)2 dt = π/2 − 4 + π = −4
0 2

In this case, Ex+y 6= Ex−y 6= Ex + Ey . Hence, we cannot generalize the conclu-

sions observed in parts (a) and (b).
R2
1.1-4. Px = 41 −2 (t3 )2 dt = 64/7
1-4

1
R2
(a) P−x = 4 −2
(−t3 )2 dt = 64/7
1
R2
(b) P2x = 4 −2
(2t3 )2 dt = 4(64/7) = 256/7
1
R2
(c) Pcx = 4 −2
(ct3 )2 dt = 64c2 /7

Comments: Changing the sign of a signal does not affect its power. Multiplying a
signal by a constant c increases the power by a factor c2 .

1.1-5. (a) Power of a sinusoid of amplitude C is C 2 /2 [Eq. (1.4a)] regardless of its frequency
(ω 6= 0) and phase. Therefore, in this case P = 52 + (10)2 /2 = 75.
(b) Power of a sum of sinusoids is equal to the sum of the powers of the sinusoids
2
(16)2
[Eq. (1.4b)]. Therefore, in this case P = (10)
2 + 2 = 178.
(c) (10 + 2 sin 3t) cos 10t = 10 cos 10t + sin 13t − sin 3t. Hence from Eq. (1.4b)
2
P = (10) 1 1
2 + 2 + 2 = 51.
(5)2 2
(d) 10 cos 5t cos 10t = 5(cos 5t+cos 15t. Hence from Eq. (1.4b) P = 2 + (5)
2 = 25.
2 2
(e) 10 sin 5t cos 10t = 5(sin 15t−sin 5t. Hence from Eq. (1.4b) P = (5) 2 + 2
(−5)
= 25.
jαt 1
£ j(α+ω0 )t j(α−ω0 )t
¤
(f) e cos ω0 t = 2 e +e . Using the result in Prob. 1.1-5, we obtain
P = (1/4) + (1/4) = 1/2.
 2A
 T t 0 ≤ t < T2 R T /2 ¡ 2A ¢2
1-5 T
Next, Px = T1 0
1.1-6. First, x(t) =

0 2 ≤t<T . T t dt =
x(t + T ) ∀t
R ¯T /2
4A2 T /2 2 4A2 t3 ¯ 2
T3 A2
T 2 0
t dt = T 2 3¯ = 4AT 2 3(8) = 6 . Since power is finite, energy must be
0
infinite. Thus,
A2
Px = and Ex = ∞.
6

24

2
 R∞ 2 R∞
1.1-7. (a)
1-6 i. By definition, E [T x1 (t)] = t=−∞ (T x1 (t)) dt = t=−∞ T 2 x21 (t)dt =
R ∞
T 2 t=−∞ x21 (t)dt = T 2 E [x1 (t)].

E [T x1 (t)] = T 2 E [x1 (t)] .

R∞ 2
ii. By definition, E [x1 (t − T )] = t=−∞ (x1 (t − T )) dt. Substituting t0 = t − T
R ∞
and dt0 = dt yields t0 =∞ x21 (t0 )dt0 = E [x1 (t)].

E [x1 (t − T )] = E [x1 (t)] .

R∞ 2
iii. By definition, E [x1 (t) + x2 (t)] = (x1 (t) + x2 (t)) dt =
R∞ ¡ 2 2
¢ t=−∞
t=−∞
x 1 (t) + 2x 1 (t)x2 (t) + x2 (t) dt. However, x1 (t) and x2 (t) are
non-overlapping so their R ∞product
¡ 2 x1 (t)x 2 (t)
¢ must beR zero. Thus,
2 ∞ 2
E [x1 (t) + x2 (t)] = x 1 (t) + x2 (t) dt = t=−∞ 1
x (t)dt +
R∞ 2
t=−∞
t=−∞ 2
x (t)dt = E [x1 (t)] + E [x2 (t)].

If (x1 (t) 6= 0) ⇒ (x2 (t) = 0) and (x2 (t) 6= 0) ⇒ (x1 (t) = 0),

Then, E [x1 (t) + x2 (t)] = E [x1 (t)] + E [x2 (t)] .

R∞
iv. By definition, E [x1 (T t)] = t=−∞ x21 (T t)dt.
First, consider
R ∞the case T > 0. Substituting
R t0 = T t and dt0 = T dt yields
0 ∞ E[x1 (t)]
E [x1 (T t)] = t0 =−∞ x21 (t0 ) dt 1 2 0
T = T t0 =−∞ x1 (t )dt =
0
T = E[x|T1 (t)]
| .
0 0
Next, considerR −∞the case T < 0. Substituting
R t = T t and dt = T dt yields
0
−1 ∞ E[x1 (t)]
E [x1 (T t)] = t0 =∞ x21 (t0 ) dt
T = T x
t0 =−∞ 1
2 0
(t )dt 0
= −T . For T < 0, we
E[x1 (t)]
know T = −|T |. Making this substitution yields E [x1 (T t)] = |T | .
Since energy is the same whether T < 0 or T > 0, we know

E [x1 (t)]
E [x1 (T t)] = .
|T |

(b) To begin, notice that signal y(t) = t(u(t)−u(t−1)) has energy equal to E [y(t)] =
R1 2
0
t dt = 1/3.
To determine E [x(t)], consider dividing x(t) into three non-overlapping pieces:
a first piece xa (t) from (−2 ≤ t < −1), a second piece xb (t) from (−1 ≤ t < 0),
and a third piece xc (t) from (0 ≤ t < 3). Since the pieces are non-overlapping,
the total energy E [x(t)] = E [xa (t)] + E [xa (t)] + E [xa (t)].
Using the properties of energy, we know that shifting or reflecting a signal does not
affect its energy. Notice that y(t/3) is the same as a flipped and shifted version of
xc (t). Thus, E [xc (t)] = E [y(t/3)] = 3(1/3) = 1. Also, it is possible to combine
xa (t) with a flipped and shifted version of xb (t) to equal a flipped and shifted
version of 2y(t/2). Thus, E [xa (t) + xb (t)] = E [2y(t/2)] = 4(2)(1/3) = 8/3.
Thus,
E [x(t)] = 11/3.
1.1-8. (a)
Z T /2 Z T /2 n X
X n
1 ∗ 1
Px = lim x(t)x (t) dt = lim Dk D∗ r ej(ωk −ωr )t dt
T →∞ T −T /2 T →∞ T −T /2 k=m r=m

The integrals of the cross-product terms (when k 6= r) are finite because the

25

3
 integrands are periodic signals (made up of sinusoids). These terms, when divided
by T → ∞, yield zero. The remaining terms (k = r) yield
Z T /2 n
X n
X
1
Px = lim |Dk |2 dt = |Dk |2
T →∞ T −T /2 k=m k=m

(b) i.

x(t) = 5 + 10 cos(100t + π/3)

π π
= 5 + 5ej(100t+ 3 ) + 5e−j(100t+ 3 )
= 5 + 5ejπ/3 ej100t + 5e−jπ/3 e−j100t

Hence,
Px = 52 + |5ejπ/3 |2 + |5e−jπ/3 |2 = 25 + 25 + 25 = 75.
Thought of another way, note that D0 = 5, D±1 = 5 and thus Px = 52 + 52 +
52 = 75.
ii.

x(t) = 10 cos(100t + π/3) + 16 sin(150t + π/5)

= 5ejπ/3 ej100t + 5e−jπ/3 e−j100t − j8ejπ/5 ej150t + j8e−jπ/5 e−j150t

Hence,

Px = |5ejπ/3 |2 +|5e−jπ/3 |2 +|−j8ejπ/5 |2 +|j8e−jπ/5 |2 = 25+25+64+64 = 178.

Thought of another way, note that D±1 = 5 and D±2 = 8. Hence, Px =

52 + 52 + 82 + 82 = 178.
iii. (10 + 2 sin 3t) cos 10t = 10 cos 10t + sin 13t − sin 3t. In this case, D±1 = 5
, D±2 = 0.5 and D±3 = 0.5 . Hence, P = 52 + 52 + (0.5)2 + (0.5)2 + (0.5)2 +
(0.5)2 = 51
iv. 10 cos 5t cos 10t = 5(cos 5t + cos 15t). In this case, D±1 = 2.5 and D±2 =
2.5. Hence, P = (2.5)2 + (2.5)2 + (2.5)2 + (2.5)2 = 25
v. 10 sin 5t cos 10t = 5(sin 15t−sin 5t). In this case, D±1 = 2.5 and D±2 = 2.5.
Hence, P = (2.5)2 + (2.5)2 + (2.5)2 + (2.5)2 = 25
£ ¤
vi. ejαt cos ω0 t = 21 ej(α+ω0 )t + ej(α−ω0 )t . In this case, D±1 = 0.5 . Hence,
P = (1/2)2 + (1/2)2 = 1/2.
1-7
1.1-9. First, notice that x(t) = x2 (t) and that the area of each pulse is one. Since x(t) has an
infinite number of pulses, the corresponding energy must alsoPbe infinite. To compute
N
the power, notice that N pulses requires an interval of width i=0 2(i+1) = N 2 +3N .
As N → ∞, power is computed by the ratio of area to width, or P = limN →∞ N 2N +3N =
0. Thus,
P = 0 and E = ∞.
1-8
1.2-1. Refer to Figure S1.2-1.
S1-8.
1-9
1.2-2. Refer to Figure S1.2-2.
S1-9.
1-10
1.2-3. (a) x1 (t) can be formed by shifting x(t) to the left by 1 plus a time-inverted version
of x(t) shifted to left by 1. Thus,

x1 (t) = x(t + 1) + x(−t + 1) = x(t + 1) + x(1 − t).

26

4
 £ ¤
tiplying it by 4/3 to obtain 34 x1 (t/2) = 34 x( t+2 2−t
2 ) + x( 2 ) . From this, we
subtract a rectangular pedestal of height 1/3 and width 4. This is obtained
by£ time-expanding¤ x2 (t) by 2 and multiplying it by 1/3 to yield 31 x2 (t/2) =
1 t+2 2−t
3 x( 4 ) + x( 4 ) . Hence,
· ¸ · ¸
4 t+2 2−t 1 t+2 2−t
x4 (t) = x( ) + x( ) − x( ) + x( ) .
3 2 2 3 4 4

(e) x5 (t) is a sum of three components: (i) x2 (t) time-compressed by a factor 2, (ii)
x(t) left-shifted by 1.5, and (iii) x(t) time-inverted and then right shifted by 1.5.
Hence,

x5 (t) = x(t + 0.5) + x(0.5 − t) + x(t + 1.5) + x(1.5 − t).

1.2-4.
1-11
Z ∞ Z ∞
2
E−x = [−x(t)] dt = x2 (t) dt = Ex
−∞ −∞
Z ∞ Z ∞
Ex(−t) = [x(−t)]2 dt = x2 (x) dx = Ex
−∞ −∞
Z ∞ Z ∞
Ex(t−T ) = [x(t − T )]2 dt = x2 (x) dx = Ex ,
−∞ −∞
Z ∞ Z
1 ∞ 2
2
Ex(at) = [x(at)] dt = x (x) dx = Ex /a
−∞ a −∞
Z ∞ Z
2 1 ∞ 2
Ex(at−b) = [x(at − b)] dt = x (x) dx = Ex /a,
−∞ a −∞
Z ∞ Z ∞
Ex(t/a) = [x(t/a)]2 dt = a x2 (x) dt = aEx
−∞ −∞
Z ∞ Z ∞
Eax(t) = [ax(t)]2 dt = a2 x2 (t) dt = a2 Ex
−∞ −∞

Comment: Multiplying a signal by constant a increases the signal energy by a factor

a2 .
1.2-5. (a) Calling y(t) = 2x(−3t + 1) = t(u(−t − 1) − u(−t + 1)), MATLAB is used to sketch
1-12
y(t).
>> t = [-1.5:.001:1.5]; y = inline(’t.*((t<=-1)-(t<=1))’);
>> plot(t,y(t),’k-’); axis([-1.5 1.5 -1.1 1.1]);
>> xlabel(’t’); ylabel(’2x(-3t+1)’);
(b) Since y(t) = 2x(−3t+1), 0.5∗y(−t/3+1/3) = 0.5(2)x(−3(−t/3+1/3)+1) = x(t).
MATLAB is used to sketch x(t).
>> y = inline(’t.*((t<=-1)-(t<=1))’);
>> t = [-3:.001:5]; x = 0.5*y(-t/3+1/3);
>> plot(t,x,’k-’); axis([-3 5 -0.6 0.6]);
>> xlabel(’t’); ylabel(’x(t)’);

1.2-6. MATLAB is used to compute each sketch. Notice that the unit step is in the exponent
1-13
of the function x(t).

28

6
 1

0.8

0.6

0.4

0.2

2x(−3t+1)
0

−0.2

−0.4

−0.6

−0.8

−1

−1.5 −1 −0.5 0 0.5 1 1.5

t

Figure S1.2-5a:
S1-12a Plot of 2x(−3t + 1).

0.5

0.4

0.3

0.2

0.1
x(t)

−0.1

−0.2

−0.3

−0.4

−0.5

−3 −2 −1 0 1 2 3 4 5
t

Figure S1.2-5b:
S1-12b Plot of x(t) = 0.5y(−t/3 + 1/3).

(a) >> t = [-1:.001:1];

>> x = inline(’2.^(-t.*(t>=0))’);
>> plot(t,x(t),’k’); axis([-1 1 0 1.1]);
>> xlabel(’t’); ylabel(’x(t)’);
(b) >> plot(t,0.5*x(1-2*t),’k’); axis([-1 1 0 1.1]);
>> xlabel(’t’); ylabel(’y(t)’);
1.3-1. (a) False. Figure 1.11b is an example of a signal that is continuous-time but digital.
1-14
(b) False. Figure 1.11c is discrete-time but analog.
(c) False. e−t is neither an energy nor a power signal.
(d) False. e−t u(t) has infinite duration but is an energy signal.
(e) False. u(t) is a power signal that is causal.
(f) True. A periodic signal, by definition, exists for all t.
1-15
1.3-2. (a) True. Every bounded periodic signal is a power signal.
(b) False. Signals with bounded power are not necessarily periodic. For example,
x(t) = cos(t)u(t) is non-periodic but has a bounded power of Px = 0.25.

29

7
 1

0.8

0.6

x(t)
0.4

0.2

0
−1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1
t

S1-13a Plot of x(t) = 2−tu(t) .

Figure S1.2-6a:

0.8

0.6
y(t)

0.4

0.2

0
−1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1
t

Figure S1.2-6b:
S1-13b Plot of y(t) = 0.5x(1 − 2t).

E
(c) True. If an energy signal x(t) has energy E, then the energy of x(at) is a (a real
and positive).
(d) False. If a power signal x(t) has power P , then the power of x(at) is generally not
P
a . A counter-example provides a simple proof. Consider the case of x(t) = u(t),
which has P = 0.5. Letting a = 2, x(at) = x(2t) = u(2t) = u(t), which still has
power P = 0.5 and not P/a = P/2.
1.3-3. (a) For periodicity, x1 (t) = cos(t) = cos(t + T1 ) = x1 (t + T1 ). Since cosine is a
1-16
2π-periodic function, T1 = 2π. Similarly, x2 (t) = sin(πt) = sin(πt + πT2 ) =
sin(π(t + T2 )) = x2 (t + T2 ). Thus, πT2 = 2πk. The smallest possible value is
T2 = 2. Thus,
T1 = 2π and T2 = 2.
(b) Periodicity requires x3 (t) = x3 (t + T3 ) or cos(t) + sin(πt) = cos(t + T3 ) + sin(πt +
πT3 ). This requires T3 = 2πk1 and πT3 = 2πk2 for some integers k1 and k2 .
Combining, periodicity thus requires T3 = 2πk1 = 2k2 or π = k1 /k2 . However, π
is irrational. Thus, no suitable k1 and k2 exist, and x3 (t) cannot be periodic.
(c)

30

8
 1
R 2π 2
Px1 = 2π ³0 cos (t)dt ´
1 2π
= 2π 0.5(t + sin(t) cos(t))| t=0
1 1 1
= 2πR2 2π = 2
1 2 2
Px2 = 2 ³0 sin (πt)dt ¯2 ´
= 1 1
(πt − sin(πt) cos(πt)) ¯
2 2π t=0 .
1 1 1
= 2 2π 2π = 2 R
1 T 2
Px3 = limt→∞ 2T −T
(cos(t) + sin(πt)) dt
R T ¡ ¢
= limt→∞ 2T 1
−T
cos2 (t) + sin2 (t) + cos(t) sin(πt) dt
1
R T
= Px1 + P x2 + limt→∞ 2T −T
0.5 (sin(πt − t) + sin(πt + t)) dt
= Px1 + P x2 + 0 = 1
Thus,
1
Px1 = Px2 = and Px3 = 1.
2
1-17
1.3-4. No, f (t) = sin(ωt) is not guaranteed to be a periodic function for an arbitrary constant
ω. Specifically, if ω is purely imaginary then f (t) is in the form of hyperbolic sine,
which is not a periodic function. For example, if ω =  then f (t) =  sinh(t). Only
when ω is constrained to be real will f (t) be periodic.
R∞ R∞
1.3-5. (a) Ey1 = −∞ y12 (t)dt = −∞ 91 x2 (2t)dt. Performing the change of variable t0 = 2t
1-18
R∞ 0
yields −∞ 91 x2 (t0 ) dt2 = E18x . Thus,

Ex 1.0417
Ey1 = ≈ = 0.0579.
18 18

(b) Since y2 (t) is just a (Ty2 = 4)-periodic replication of x(t), the power is easily
obtained as
Ex Ex
Py2 = = ≈= 0.2604.
Ty2 4
R R
(c) Notice, Ty3 = Ty2 /2 = 2. Thus, Py3 = 21 Ty y32 (t)dt = 21 Ty 91 y2 (2t)dt.
3
R 3
0
Performing the change of variable t0 = 2t yields Py3 = 21 Ty 91 y2 (t0 ) dt2 =
1
R4 0 0 Ex
2

36 0 x(t )dt = 36 . Thus,

Ex
Py3 = ≈ 0.0289.
36
1-19
1.3-6. For all parts, y1 (t) = y2 (t) = t2 over 0 ≤ t ≤ 1.

(a) To ensure y1 (t) is even, y1 (t) = t2 over −1 ≤ t ≤ 0. ½ Since2 y1 (t) is (T1 = 2)-
t −1 ≤ t ≤ 1
periodic, y1 (t) = y1 (t + 2) for all t. Thus, y1 (t) = .
y1 (t + 2) ∀t
R1 5¯
¯ 1
Py1 = T11 −1 (t2 )2 dt = 0.5 t5 ¯ = 1/5. Thus,
t=−1

Py1 = 1/5.

A sketch of y1 (t) over −3 ≤ t ≤ 3 is created using MATLAB.

>> t = [-3:.001:3]; mt = mod(t,2);
>> y_1 = (mt<=1).*(mt.^2) + (mt>1).*((mt-2).^2);
>> plot(t,y_1,’k’); xlabel(’t’); ylabel(’y_1(t)’); axis tight;

31

9
 1

0.9

0.8

0.7

0.6

y (t)
0.5

1
0.4

0.3

0.2

0.1

0
−3 −2 −1 0 1 2 3
t

Figure S1.3-6a:
S1-19a Plot of y1 (t)

(b) Let 

 k 1 ≤ t < 1.5

t2 0≤t<1
y2 (t) = .

 −y2 (−t) ∀t

y2 (t + 3) ∀t
With this form, y2 (t) is odd and (T2 = 3)-periodic. The
³ constant k is determined
¯1 ´
by constraining the power to be unity, Py2 = 1 = 31 k 2 + 52 t5 ¯t=0 . Solving for
p
k yields k 2 = 3 − 2/5 = 13/5 or k = 13/5. Thus,
 p

 13/5 1 ≤ t < 1.5

t2 0≤t<1
y2 (t) = .

 −y 2 (−t) ∀t

y2 (t + 3) ∀t

A sketch of y2 (t) over −3 ≤ t ≤ 3 is created using MATLAB.

>> t = [-3:.001:3]; mt = mod(t,3);
>> y_2 = (mt<1).*(mt.^2)-(mt>=2).*((mt-3).^2)+...
((mt>=1)&(mt<1.5))*sqrt(13/5)-...
((mt>=1.5)&(mt<2))*sqrt(13/5);
>> plot(t,y_2,’k’); xlabel(’t’); ylabel(’y_2(t)’); axis([-3 3 -2 2]);
(c) Define y3 (t) = y1 (t) + y2 (t). To be periodic, y3 (t) must equal y3 (t + T3 ) for some
value T3 . This implies that y1 (t) = y1 (t + T3 ) and y2 = y2 (t + T3 ). Since y1 (t)
is (T1 = 2)-periodic, T3 must be an integer multiple of T1 . Similarly, since y2 (t)
is (T2 = 3)-periodic, T3 must be an integer multiple of T2 . Thus, periodicity of
y3 (t) requires T3 = T1 k1 = 2k1 = T2 k2 = 3k2 , which is satisfied letting k1 = 3
and k2 = 2. Thus,
y3 (t) is periodic with T3 = 6.
R ¡ ¢
(d) Noting y3 (t)y3∗ (t) = y12 (t) + y22 (t), Py3 = T13 T3 y12 (t) + y22 (t) dt = Py1 + Py2 .
Thus,
1 6
P y3 = 1 + = .
5 5
1.4-1. Refer to Figure S1-20.
1-20 S1.4-1.

32

10
 R∞
(b) Exe = −∞ x2e (t)dt. Because e−2t u(t) and e2t u(−t) are disjoint in time, the
cross-product term in x2e (t) is zero. Hence,
Z ∞ ·Z ∞ Z ¸
0
1 1
Exe = x2e (t)dt = e −4t
dt + e dt = . 4t
−∞ 4 0 −∞ 8

Using a similar argument, we have

1
Exo = .
8
Also, Z ∞
1
Ex = e−4t dt = .
0 4
Hence,
Ex = Exe + Exo .
(c) To generalize this result, we first consider causal x(t). In this case, x(t) and x(−t)
are disjoint. Moreover, energy or x(t) is identical to that of x(−t). Hence,
·Z ∞ Z 0 ¸
1 1
Exe = |x(t)|2 dt + |x(−t)|2 dt = Ex .
4 0 −∞ 2

Using a similar argument, it follows that Exo = 21 Ex . Hence, for causal signals,

Ex = Exe + Exo

Identical arguments hold for anti-causal signals. Thus, for anti-causal signal x(t)

Ex = Exe + Exo

Now, every signal can be expressed as a sum of a causal and an anti-causal

signal. Also, the signal energy is equal to the sum of energies of the causal and
the anti-causal components. Hence, it follows that for a general case

Ex = Exe + Exo

1.5-3. (a)
1-32
1
xe (t)xo (t) = [x(t) + x(−t)][x(t) − x(−t)]
4
1
= [|x(t)|2 − |x(−t)|2 ]
4
Since the areas under |x(t)|2 and |x(−t)|2 are identical, it follows that
Z ∞
xe (t)xo (t)dt = 0
−∞

(b)
Z ∞ Z ∞ Z ∞
1 1
xe (t)dt = x(t)dt + x(−t)dt
−∞ 2 −∞ 2 −∞

37

15
 Because the areas under x(t) and x(−t) are identical, it follows that
Z ∞ Z ∞
xe (t)dt = x(t)dt.
−∞ −∞

1-33
1.5-4. xo (t) = 0.5(x(t) − x(−t)) = 0.5(sin(πt)u(t) − sin(−πt)u(−t)) = 0.5 sin(πt)(u(t) +
u(−t)). Since sin(0) = 0, this reduces to xo (t) = 0.5 sin(πt), which is a (T = 2)-
periodic signal. Therefore,

xo (t) = 0.5 sin(πt) is a periodic signal.

1-34
1.5-5. xe (t) = 0.5(x(t) + x(−t)) = 0.5(cos(πt)u(t) + cos(−πt)u(−t)) = 0.5 cos(πt)(u(t) +
½
0.5 cos(πt) t 6= 0
u(−t)). Written another way, xe (t) = . Since there exists no
1 t=0
T 6= 0 such that xe (t + T ) = xe (t),

xe (t) is not a periodic function.

It is worth pointing out that sometimes the unit step is defined as u(t) =

 1 t>0
0.5 t = 0 . Using this alternate definition, xe (t) is periodic.

0 t<0
1-35
1.5-6. (a) Using the figure, x(t) = (t+1)(u(t+1)−u(t))+(−t+1)(u(t)−u(t−1)). MATLAB
¡ ¢
is used to plot v(t) = 3x − 21 (t + 1) .
>> x = inline(’(t+1).*((t>=-1)&(t<0))+(-t+1).*((t>=0)&(t<1))’);
>> t = [-5:.001:5]; v = 3*x(-0.5*(t+1));
>> plot(t,v,’k-’); xlabel(’t’); ylabel(’v(t)’);

2.5

2
v(t)

1.5

0.5

0
−5 −4 −3 −2 −1 0 1 2 3 4 5
t
¡¢
S1-35a Plot of v(t) = 3x − 21 (t + 1) .
Figure S1.5-6a:

(b) Since v(t) is finite duration, Pv = 0. Signal is unaffected by shifting, so v(t) is

shifted to start at t = 0. By symmetry, the energy of the first ¯half is equal to
R 2 ¡ ¢2 3 ¯t=2
the energy of the second half. Thus, Ev = 2 0 23 t dt = 2 49 t3 ¯ = 24
2 = 12.
t=0
Thus,
Ev = 12 and Pv = 0.

38

16
(c) Using 1.5-6a,
1-35a v(t) = (1.5t+4.5)(u(t+3)−u(t+1))+(−1.5t+1.5)(u(t+1)−u(t−1)).
MATLAB is used to determine and plot ve (t).
>> v_str = [’(1.5*t+4.5).*((t>=-3)&(t<-1))+’,...
’(-1.5*t+1.5).*((t>=-1)&(t<1))’];
>> v = inline(v_str); t = [-4:.001:4]; v_e = 0.5*(v(t)+v(-t));
>> plot(t,v_e,’k’); xlabel(’t’); ylabel(’v_e(t)’);
>> axis([-4 4 -.25 1.75]);

1.6

1.4

1.2

0.8
ve(t)

0.6

0.4

0.2

−0.2
−4 −3 −2 −1 0 1 2 3 4
t

Figure S1-35c
S1.5-6c: Plot of ve (t).

Thus, 

 3t/4 + 9/4 −3 ≤ t < −1

1.5 −1 ≤ t < 1
ve (t) = .

 −3t/4 + 9/4 1≤t<3

0 otherwise
(d) Using 1.5-6a,
1-35a, v(t) = (1.5t+4.5)(u(t+3)−u(t+1))+(−1.5t+1.5)(u(t+1)−u(t−1)).
MATLAB is used to create the four desired plots.
>> v_str = [’(1.5*t+4.5).*((t>=-3)&(t<-1))+’,...
’(-1.5*t+1.5).*((t>=-1)&(t<1))’];
>> v = inline(v_str);
>> t = [-6:.001:1]; a = 2; b = 3; ax = [-6 1 -.5 9.5];
>> subplot(221); plot(t,v(a*t+b),’k-’); grid;
>> xlabel(’t’); ylabel(’v(at+b)’); axis(ax);
>> subplot(222); plot(t,v(a*t)+b,’k-’); grid;
>> xlabel(’t’); ylabel(’v(at)+b’); axis(ax);
>> subplot(223); plot(t,a*v(t+b),’k-’); grid;
>> xlabel(’t’); ylabel(’av(t+b)’); axis(ax);
>> subplot(224); plot(t,a*v(t)+b,’k-’); grid;
>> xlabel(’t’); ylabel(’av(t)+b’); axis(ax);
(e) Using 1.5-6a,
1-35a v(t) = (1.5t+4.5)(u(t+3)−u(t+1))+(−1.5t+1.5)(u(t+1)−u(t−1)).
MATLAB is used to create the four desired plots.
>> v_str = [’(1.5*t+4.5).*((t>=-3)&(t<-1))+’,...
’(-1.5*t+1.5).*((t>=-1)&(t<1))’];
>> v = inline(v_str);
>> t = [-3:.001:3]; a = -3; b = -2; ax = [-3 3 -11.5 3.5];

39

17
 8 8

6 6

v(at+b)

v(at)+b
4 4

2 2

0 0
−6 −4 −2 0 −6 −4 −2 0
t t

8 8

6 6

av(t+b)

av(t)+b
4 4

2 2

0 0
−6 −4 −2 0 −6 −4 −2 0
t t

Figure S1-35d
S1.5-6d: Plots of v(2t + 3), v(2t) + 3, 2v(t + 3), and 2v(t) + 3.

>> subplot(221); plot(t,v(a*t+b),’k-’); grid;

>> xlabel(’t’); ylabel(’v(at+b)’); axis(ax);
>> subplot(222); plot(t,v(a*t)+b,’k-’); grid;
>> xlabel(’t’); ylabel(’v(at)+b’); axis(ax);
>> subplot(223); plot(t,a*v(t+b),’k-’); grid;
>> xlabel(’t’); ylabel(’av(t+b)’); axis(ax);
>> subplot(224); plot(t,a*v(t)+b,’k-’); grid;
>> xlabel(’t’); ylabel(’av(t)+b’); axis(ax);

2 2
0 0
−2 −2
v(at+b)

v(at)+b

−4 −4
−6 −6
−8 −8
−10 −10

−3 −2 −1 0 1 2 3 −3 −2 −1 0 1 2 3
t t

2 2
0 0
−2 −2
av(t+b)

av(t)+b

−4 −4
−6 −6
−8 −8
−10 −10

−3 −2 −1 0 1 2 3 −3 −2 −1 0 1 2 3
t t

Figure S1.5-6e:
S1-35e Plots of v(−3t − 2), v(−3t) − 2, −3v(t − 2), and −3v(t) − 2.

1-36 (a) Using the figure, y(t) = t(u(t) − u(t − 1)) + (u(t − 1) − u(t − 2)). MATLAB is
1.5-7.
used to plot yo (t) = y(t)−y(−t)
2 .
>> y = inline(’t.*((t>=0)&(t<1))+((t>=1)&(t<2))’);
>> t = [-5:.001:5]; y_o = (y(t)-y(-t))/2;
>> plot(t,y_o,’k-’); xlabel(’t’); ylabel(’y_o(t)’); axis([-5 5 -.6 .6]);

40

18
 0.5

0.4

0.3

0.2

0.1

yo(t)
0

−0.1

−0.2

−0.3

−0.4

−0.5

−5 −4 −3 −2 −1 0 1 2 3 4 5
t

y(t)−y(−t)
Figure S1.5-7a:
S1-36a Plot of yo (t) =
2 .

Thus, 

 −1/2 −2 ≤ t < −1

t/2 −1 ≤ t < 1
yo (t) = .

 1/2 1≤t<2

0 otherwise

(b) Since y(t) = 0.2x(−2t−3), 5y(−0.5t−1.5) = 5(0.2)x(−2(−0.5t−1.5)−3) = x(t).

MATLAB is used to sketch x(t).
>> y = inline(’t.*((t>=0)&(t<1))+((t>=1)&(t<2))’);
>> t = [-8:.001:0]; x = 5*y(-0.5*t-1.5);
>> plot(t,x,’k-’); xlabel(’t’); ylabel(’x(t)’); axis([-8 0 -.5 5.5]);

3
x(t)

−8 −7 −6 −5 −4 −3 −2 −1 0
t

Figure S1.5-7b:
S1-36b Plot of x(t) = 5y(−0.5t − 1.5).

Thus, 
 5 −7 ≤ t < −5
x(t) = −5(t + 3)/2 −5 ≤ t < −3 .

0 otherwise

1.5-8.
1-37 Let the graphed signal be named y(t).

41

19
(a) Since y(t) = −0.5x(−3t+2), −2y(−t/3+2/3) = −2(−0.5)x(−3(−t/3+2/3)+2) =
x(t). MATLAB is used to sketch x(t).
>> y = inline(’((t>=-1)&(t<0))+(-t+1).*((t>=0)&(t<1))’);
>> t = [-2:.001:6]; x = -2*y(-t/3+2/3);
>> plot(t,x,’k-’); xlabel(’t’); ylabel(’x(t)’); axis([-2 6 -2.5 0.5]);

0.5

−0.5
x(t)

−1

−1.5

−2

−2.5
−2 −1 0 1 2 3 4 5 6
t

Figure S1.5-8a:
S1-37a: Plot of x(t) = −2y(−t/3 + 2/3).

Thus, 
 −2(t + 1)/3 −1 ≤ t < 2
x(t) = −2 2≤t<5 .

0 otherwise

(b) The even portion of x(t) is xe (t) = 0.5(x(t) + x(−t)).

>> x = inline(’-2*(t+1)/3.*((t>=-1)&(t<2))-2*((t>=2)&(t<5))’);
>> t = [-6:.001:6]; x_e = 0.5*(x(t)+x(-t));
>> plot(t,x_e,’k-’); xlabel(’t’); ylabel(’x_e(t)’); axis([-6 6 -1.5 0.5]);

0.5

0
xe(t)

−0.5

−1

−1.5
−6 −4 −2 0 2 4 6
t

Figure S1.5-8b:
S1-37b Plot of xe (t).

42

20
 Thus, 

 −1 2 ≤ |t| < 5

(−|t| − 1)/3 1 ≤ |t| < 2
xe (t) = .

 −2/3 |t| < 1

0 otherwise

(c) The odd portion of x(t) is xo (t) = 0.5(x(t) − x(−t)).

>> x = inline(’-2*(t+1)/3.*((t>=-1)&(t<2))-2*((t>=2)&(t<5))’);
>> t = [-6:.001:6]; x_o = 0.5*(x(t)-x(-t));
>> plot(t,x_o,’k-’); xlabel(’t’); ylabel(’x_e(t)’); axis([-6 6 -1.5 1.5]);

1.5

0.5
xe(t)

−0.5

−1

−1.5
−6 −4 −2 0 2 4 6
t

Figure S1.5-8c:
S1-37c: Plot of xo (t).

Thus, 

 1 −5 ≤ t < −2



 (−t + 1)/3 −2 ≤ t < −1

−2t/3 −1 ≤ t < 1
xo (t) = .

 −(t + 1)/3 1≤t<2



 −1 2≤t<5

0 otherwise
∗
1.5-9. Notice, wcs
1-38 (−t) = 0.5(w(−t) + w∗ (t))∗ = 0.5(w∗ (−t) + w(t)) = wcs (t). In Cartesian
∗
form, this becomes wcs (−t) = x(−t) − y(−t) = x(t) + y(t) = wcs (t). Equating
the real portions yields x(−t) = x(t), and equating the imaginary portions yields
−y(−t) = y(t). Thus, by definition, the real portion of wcs (t) is even and the imaginary
portion of wcs (t) is odd.
1-39
1.5-10. Notice, −wca∗
(−t) = −0.5(w(−t) − w∗ (t))∗ = 0.5(−w∗ (−t) + w(t)) = wca (t). In
∗
Cartesian form, this becomes −wca (−t) = −x(−t) + y(−t) = x(t) + y(t) = wca (t).
Equating the real portions yields −x(−t) = x(t), and equating the imaginary portions
yields y(−t) = y(t). Thus, by definition, the real portion of wca (t) is odd and the
imaginary portion of wca (t) is even.
1-40
1.5-11. Complex signal w(t) is defined over (0 ≤ t ≤ 1).
(a) Assigning certain properties to w(t) allows us to plot w(t) over (−1 ≤ t ≤ 1).
i. If w(t) is even, w(t) = w(−t), it is even in both the real and imaginary
components. Thus, the graph folds back on itself and appears unchanged.

43

21
 Consider, for example, point (2,1), which now corresponds to both t = 1 and
t = −1.
ii. If w(t) is odd, w(t) = −w(−t), it is odd in both the real and imaginary
components. Thus, the graph reflects about both the real and imaginary
axes.
iii. If w(t) is conjugate symmetric, w(t) = w∗ (−t), it is even in the real compo-
nent and odd in the imaginary component. Thus, the graph reflects about
the real axis.
iv. If w(t) is conjugate antisymmetric, w(t) = −w∗ (−t), it is odd in the real
component and even in the imaginary component. Thus, the graph reflects
about the imaginary axis.

w(t) is Even w(t) is Odd

1.5 1.5

1 1

0.5 0.5
t=−1 t=−1
Im

Im
0 t=0 0 t=0
t=1 t=1
−0.5 −0.5

−1 −1

−1.5 −1.5
−2 −1 0 1 2 −2 −1 0 1 2
Re Re

w(t) is Conjugate Symmetric w(t) is Conjugate Antisymmetric

1.5 1.5

1 1

0.5 0.5
t=−1 t=−1
Im

Im

0 t=0 0 t=0
t=1 t=1
−0.5 −0.5

−1 −1

−1.5 −1.5
−2 −1 0 1 2 −2 −1 0 1 2
Re Re

Figure S1.5-11a:
S1-40a Plots of w(t).

(b) Since w(t) is only given over (0 ≤ t ≤ 1), w(3t) can be determined only for
(0 ≤ t ≤ 1/3). Since the function does not change, only the time at which it
occurs, the complex-plane graph of w(3t) looks identical to the original complex-
plane graph of w(t) with the exception that the points are assigned different
times. For example, point (2,1) occurs now at t = 1/3.
1-41 (a) We know x(t) = t2 (1 + ) over (1 ≤ t ≤ 2). Since x(t) is skew-Hermitian,
1.5-12.
x(t) = −x∗ (−t) and thus x(t) = −t2 (1 − ) over (−2 ≤ t ≤ −1). To minimize
energy, x(t) is set to zero everywhere else. Thus,
 2
 t (1 + ) 1≤t≤2
x(t) = −t2 (1 − ) −2 ≤ t ≤ −1 .

0 otherwise

(b) MATLAB is used to sketch y(t) = Real {x(t)}.

>> x_str = [’(t.^2*(1+j)).*((t>=1)&(t<=2))+’,...
’(-t.^2*(1-j)).*((t>=-2)&(t<=-1))’];
>> x = inline(x_str);
>> t = [-2.5:.001:2.5]; plot(t,real(x(t)),’k’);
>> xlabel(’t’); ylabel(’y(t)’); axis([-2.5 2.5 -4.5 4.5]);
(c) MATLAB is used to sketch z(t) = Real {x(−2t + 1)}.

44

22
 4

y(t)
0

−1

−2

−3

−4

−2.5 −2 −1.5 −1 −0.5 0 0.5 1 1.5 2 2.5

t

Figure S1.5-12b:
S1-41b Plot of y(t) = Real {x(t)}.

>> x_str = [’(t.^2*(1+j)).*((t>=1)&(t<=2))+’,...

’(-t.^2*(1-j)).*((t>=-2)&(t<=-1))’];
>> x = inline(x_str);
>> t = [-2.5:.001:2.5]; plot(t,real(j*x(-2*t+1)),’k’);
>> xlabel(’t’); ylabel(’z(t)’); axis([-2.5 2.5 -4.5 4.5]);

1
z(t)

−1

−2

−3

−4

−2.5 −2 −1.5 −1 −0.5 0 0.5 1 1.5 2 2.5

t

Figure S1.5-12c:
S1-41c Plot of z(t) = Real {x(−2t + 1)}.

R2
(d) Since x(t) is finite duration, Px = 0. Using symmetry, Ex = 2 1
(t2 (1+))(t2 (1−
R2 ¯
5 ¯2
))dt = 2 1 2t4 dt = 4t5 ¯ = 54 (32 − 1) = 124
5 . Thus,
t=1

124
Ex = = 24.8 and Px = 0.
5

1.6-1. If x(t) and y(t) are the input and output, respectively, of an ideal integrator, then
1-42

ẏ(t) = x(t)

45

23
 and
Z t Z 0 Z t Z t
y(t) = x(τ ) dτ = x(τ ) dτ + x(τ ) dτ = y(0) + x(τ ) dτ
−∞ −∞ 0 |{z} 0
zero-input | {z }
zero-state
1-43
1.6-2. From Newton’s law
dv
x(t) = M
dt
and
Z t Z 0 Z t Z t
1 1 1 1
v(t) = x(τ ) dτ = x(τ ) dτ + x(τ ) dτ = v(0) + x(τ ) dτ
M −∞ M −∞ M 0 M 0

1.7-1. Only (b), (f), and (h) are linear. All the remaining are nonlinear. This can be verified
1-44
by using the procedure discussed in Example 1.9.
1-45
1.7-2. (a) The system is time-invariant because the input x(t) yields the output y(t) =
x(t − 2). Hence, if the input is x(t − T ), the output is x(t − T − 2) = y(t − T ),
which makes the system time-invariant.
(b) The system is time-varying. The input x(t) yields the output y(t) = x(−t).
Thus, the output is obtained by changing the sign of t in x(t). Therefore, when
the input is x(t − T ), the output is x(−t − T ) = x(−[t + T ]) = y(t + T ), which
represents the original output advanced by T (not delayed by T ).
(c) The system is time-varying. The input x(t) yields the output y(t) = x(at), which
is a scaled version of the input. Thus, the output is obtained by replacing t in
the input with at. Thus, if the input is x(t − T ) (x(t) delayed by T ), the output
is x(at − T ) = x(a[t − Ta ]), which is x(at) delayed by T /a (not T ). Hence the
system is time-varying.
(d) The system is time-varying. The input x(t) yields the output y(t) = tx(t). For
the input x(t−T ), the output is tx(t−T ), which is not tx(t) delayed by T . Hence
the system is time-varying.
(e) The system is time-varying. The output is a constant, given by the area under
x(t) over the interval |t| ≤ 5. Now, if x(t) is delayed by T , the output, which is
the area under the delayed x(t), is another constant. But this output is not the
same as the original output delayed by T . Hence the system is time-varying.
(f) The system is time-invariant. The input x(t) yields the output y(t), which is the
square of the second derivative of x(t). If the input is delayed by T , the output
is also delayed by T . Hence the system is time-invariant.
1.7-3. We construct the table below from the first three rows of data. Let rj denote the jth
1-46
row.

Row x(t) q1 (0) q2 (0) y(t)

r1 0 1 −1 e−t u(t)
r2 0 2 1 e−t (3t + 2)u(t)
r3 u(t) −1 −1 2u(t)
r4 = 31 (r1 + r2 ) 0 1 0 (t + 1)e−t u(t)
1
r5 = 21 (r1 + r3 ) 2 u(t) 0 −1 ( 21 e−t + 1)u(t)
1
r6 = (r4 + r5 ) 2 u(t) 1 −1 (1.5e + te−t + 1)u(t)
−t

r7 = 2(r6 + r1 ) u(t) 0 0 (e−t + 2te−t + 2)u(t)

46

24
 (c) Yes, the system is causal. Since the system is memoryless, the system cannot
depend on future values and must be causal.
(d) No, the system is not time-invariant. Since the system function depends on
the independent variable t, it is unlikely that the system is time-invariant. To
explicitly verify, let y(t) = r(t)x(t). Next, delay x(t) by τ to obtain a new
input x2 = x(t − τ ). Applying x2 (t) to the system yields y2 (t) = r(t)x2 (t) =
r(t)x(t − τ ) 6= r(t − τ )x(t − τ ) = y(t − τ ). Since, the system operator and the
time-shift operator do not commute, the system is not time-invariant.
1.7-10.
1-53 Using the sifting property, this system operation is rewritten as y(t) =
0.5 (x(t) − x(−t)).
(a) This system extracts the odd portion of the input.
(b) Yes, the system is BIBO stable. If the input is bounded, then the output is neces-
sarily bounded. That is, if |x(t)| ≤ Mx < ∞, then |y(t)| = |0.5 (x(t) − x(−t)) | ≤
0.5 (|x(t)| + | − x(−t)|) ≤ Mx < ∞.
(c) Yes, the system is linear. Let y1 (t) = 0.5(x1 (t) − x1 (−t)) and y2 (t) =
0.5(x2 (t) − x2 (−t)). Applying ax1 (t) + bx2 (t) to the system yields y(t) =
0.5 (ax1 (t) + bx2 (t) − (ax1 (−t) + bx2 (−t))) = 0.5a(x1 (t) − x1 (−t)) + 0.5b(x2 (t) −
x2 (−t)) = ay1 (t) + by2 (t).
(d) No, the system is not memoryless. For example, at time t = 1 the output
y(1) = 0.5(x(1) − x(−1)) depends on a past value of the input, x(−1).
(e) No, the system is not causal. For example, at time t = −1 the output y(−1) =
0.5(x(−1) − x(1)) depends on a future value of the input, x(1).
(f) No, the system is not time-invariant. For example, let the input be x(t) =
t(u(t + 1) − u(t − 1)). Since this input is already odd, the output is just the input,
y(t) = x(t). Shifting by a non-zero τ , x(t − τ ) is not odd, and the output is not
y(t − τ ) = x(t − τ ). Thus, the system cannot be time-invariant.
1-54 (a) No, the system is not BIBO stable. The system returns the time-delayed deriva-
1.7-11.
tive, or slope, of the input signal. A square-wave is a bounded signal which, due
to point discontinuities, has infinite slope at certain instants in time. Thus, a
bounded input may not result in a bounded output, and the system cannot be
BIBO stable.
d
(b) Yes, the system is linear. Begin assuming y1 (t) = dt x1 (t − 1) and
d
y2 (t) = dt x2 (t − 1). Applying ax1 (t) + bx2 (t) to the system yields y(t) =
d d d
dt (ax1 (t − 1) + bx2 (t − 1)) = a dt x1 (t − 1) + b dt x2 (t − 1) = ay1 (t) + by2 (t).
(c) No, the system is not memoryless. By inspection, it is clear that the system
depends on a past value of the input. For example, at t = 0, the output y(0)
depends on the time-derivative of x(−1), a past value.
(d) Yes, the system is causal. By inspection, it is clear that the system does not
depend on future values.
d
(e) Yes, the system is time-invariant. To explicitly verify, let y(t) = dt x(t − 1). Next,
delay x(t) by τ to obtain a new input x2 = x(t−τ ). Applying x2 (t) to the system
d d
yields y2 (t) = dt x2 (t) = dt x(t − 1 − τ ) = y(t − τ ). Since, the system operator
and the time-shift operator commute, the system is time-invariant. In more loose
terms, the derivative operator returns the delayed slope of a signal independent
of when that signal is applied.

49

27
 1-55 (a) Yes, the system is BIBO stable. From the definition of the system, we know that
1.7-12.
y(t) is either x(t) or 0. Correspondingly, if |x(t)| < ∞ then |y(t)| < ∞, and the
system must be BIBO stable.
(b) No, the system is not linear. Consider two signals: x1 (t) = 1 and x2 (t) =
cos(t). The
½ corresponding outputs of these individual signals is y1 (t) = 1 and
cos(t) if cos(t) > 0
y2 (t) = . However, if we create a third input x3 (t) =
0 if cos(t) ≤ 0
x1 (t) + x2 (t), the system output is y3 (t) = 1 + cos(t) 6= y1 (t) + y2 (t). Since
superposition does not apply, the system cannot be linear.
(c) Yes, the system is memoryless. By inspection, it is clear that the system only
depends on the current input.
(d) Yes, the system is causal. Since the system is memoryless, the system cannot
depend on future values and must be causal.
(e) Yes, the system is time-invariant. Consider delaying x(t) by τ to obtain a
new
½ input x2 = x(t − τ ). Applying x2 (t) to the system yields y2 (t) =
x(t − τ ) if x(t − τ ) > 0
= y(t − τ ). Since the system operator and the
0 if x(t − τ ) ≤ 0
time-shift operator commute, the system is time-invariant.
1-56
1.7-13. (a) No, Bill is not correct. The x1 (3t) term represents a compression rather than the
necessary dilation. One way to construct x2 (t) is x2 (t) = 2x1 (t/3) − x1 (t − 1).
However, this form is not unique; x2 (t) = 2x1 (t) + x1 (t − 1) + 2x1 (t − 2) also
works and may be more useful.
(b) The output y1 (t) is given for the input signal x1 (t). The expression x2 (t) =
2x1 (t)+x1 (t−1)+2x1 (t−2) forms x2 (t) from a superposition of scaled and shifted
copies of x1 (t). Since the system is linear and time invariant, the operations
of scaling, summing, and shifting commute with the system operator. Thus,
y2 (t) = 2y1 (t) + y1 (t − 1) + 2y1 (t − 2). Notice, it is not true that y2 (t) =
2y1 (t/3) − y1 (t − 1); linearity and time-invariance cannot help with the time-
scaling operation. MATLAB is used to plot y2 (t).
>> t = [-1:.001: 4]; y_1 = inline(’t.*((t>=0)&(t<1))+(t>=1)’);
>> y_2 = 2*y_1(t)+y_1(t-1)+2*y_1(t-2);
>> plot(t,y_2,’k-’); xlabel(’t’); ylabel(’y_2(t)’); axis([-1 4 -.5 5.5]);

3
y2(t)

−1 −0.5 0 0.5 1 1.5 2 2.5 3 3.5 4

t

Figure S1.7-13b:
S1-56b Plot of y2 (t) = 2y1 (t) + y1 (t − 1) + 2y1 (t − 2).

50

28
 Substituting of (1) in (2) yields

Kf
x(t) = (JD2 + BD)θ(t)
DLf + Rf
or
(JD2 + BD)(DLf + Rf )θ(t) = Kf x(t)
1.8-5.
1-61

[qi (t) − q0 (t)]4t = A4h

or
1
ḣ(t) = [qi (t) − q0 (t)] (1)
A
But
q0 (t) = Rh(t) (2)
Differentiation of (2) yields

R
q̇0 (t) = Rḣ(t) = [qi (t) − q0 (t)]
A
and µ ¶
R R
D+ q0 (t) = qi (t)
A A
or
R
(D + a)q0 (t) = aqi (t) a= (3)
A
and
a
q0 (t) = qi (t)
D+a
substituting this in (1) yields
µ ¶
1 a D
ḣ(t) = 1− qi (t) = qi (t)
A D+a A(D + a)
or
1
(D + a)h(t) = qi (t)
A
1.8-6.
1-62 (a) The order of the system is zero; there are no energy storage components such as
capacitors or inductors.
(b) Using KVL on the left loop yields x(t) = R1 y1 (t) + R2 (y1 (t) − y2 (t)) = 3y1 (t) −
2y2 (t). KVL on the middle loop yields 0 = R2 (y2 (t)−y1 (t))+R3 y2 (t)+R4 (y2 (t)−
y3 (t)) = −2y1 (t)+9y2 (t)−4y3 (t). Finally, KVL on the right loop yields R4 (y3 (t)−
y2 (t)) + (R5 + R6 )y3 (t) = −4y2 (t) + 15y3 (t). Combining together yields
    
3 −2 0 y1 (t) x(t)
 −2 9 −4   y2 (t)  =  0  .
0 −4 15 y3 (t) 0

52

30
 (c) Cramer’s rule suggests
¯ ¯
¯ 3 −2 x(t) ¯
¯ ¯
¯ −2 9 0 ¯¯
¯
¯ 0 −4 0 ¯
y3 (t) = ¯¯ ¯.
¯
¯ 3 −2 0 ¯
¯ −2 9 −4 ¯
¯ ¯
¯ 0 −4 15 ¯

MATLAB computes the denominator determinant.

>> det([3 -2 0;-2 9 -4;0 -4 15])
ans = 297
The numerator determinant is easy computed by hand as 0 + 0 + 8x(t) = 8x(t).
Thus,
8 8
y3 (t) = x(t) = (2 − | cos(t)|) u(t − 1).
297 297
1.10-1.
1-63 From Figure P1.8-2,
P1-58, we obtain

x(t) = q1 /2 + q̇1 + q2

Moreover, the capacitor voltage q1 (t) equals the voltage across the inductor, which is
1
2 q̇2 . Hence, the state equations are

q̇1 = −q1 /2 − q2 − x and q̇2 = 2q1

1-64 The capacitor current C q̇3 = 21 q̇3 is q1 − q2 . Therefore

1.10-2.

q̇3 = 2q1 − 2q2 (1)

The two loop equations are

2q1 + q̇1 + q3 = x =⇒ q̇1 = −2q1 − q3 + x (2)

1
−q3 + q̇2 + q2 = 0 =⇒ q̇2 = −3q2 + 3q3 (3)
3
Equations (1), (2) and (3) are the state equations.
For the 2Ω resistor: current is q1 , voltage is 2q1 .
For the 1H inductor: current is q1 , voltage is q̇1 = x(t) − 2q1 − q3 .
For the capacitor: current is q1 − q2 , voltage is q3 .
For the 31 H inductor: current is q2 , voltage is 31 q̇2 = −q2 + q3 .
For the 1Ω resistor: current is q2 and voltage is q2 .
At the instant t, q1 = 5, q2 = 1, q3 = 2 and x = 10. Substituting these values in the
above results yields
2Ω resistor: current 5A, voltage 10A.
1H capacitor: current 5A, voltage 10 − 10 − 2 = −2V.
The capacitor: current 5 − 1 = 4A, voltage 2V.
The 31 H inductor: current 1A, voltage −1 + 2 = 1V.
The 1Ω resistor: current 1A, voltage 1V.

53

31