Training 1
Training 1
1.1. A continuous-time signal x( t) is shown in Fig. 1-17. Sketch and label each of the following signals.
(a) x(t- 2); (b) x(2t); (c) x(t/2); (d) x(-t)
x(t)
-2-1 0 1 2 3 4 5
Fig. 1-17
3 3
- 1 0 1 2 3 4 5 6 7 - 2- 1 0 1 2 3
(a) (b)
x(t/2) x( - t)
3 3
- 1 0 1 2 3 4 5 6 7 8 9 - 5 - 4- 3- 2 - 1 0 1 2
(c) (d)
Fig. 1-18
1.2. A discrete-time signal x[n] is shown in Fig. 1-19. Sketch and label each of the following signals.
(a) x[n - 2]; (b) x[2n]; (c) x[-n]; (d) x[-n + 2]
x[n]
- 1 O 1 2 3 4 5 n
Fig. 1-19
x[n-2] x[2n]
3 3
01234567 n -1 0 1 2 3 n
(a) (b)
x[- n] x[- n + 2]
3 3
- 5 - 4 - 3- 2 - 1 0 1 n -3- 2 - 1 0 1 2 n
(c) (d)
Fig. 1-20
determine the resultant discrete-time sequence obtained by uniform sampling of x(t) with a sampling
interval of (a) 0.25 s, (b) 0.5 s, and (c) 1.0 s.
It is easier to take the graphical approach for this problem. The signalx(t) is plotted in Fig . 1-21(a). Figs. 1-2l(b) to (d)
give plots of the resultant sampled sequences obtained for the three specified sampling intervals.
- 1 0 - 4- 3- 2- 1 0 1 2 3 4 n
(a) (b)
-2 -1 0 2 n - 1 0 n
(c) (d)
Fig. 1-21
1.4. Using the discrete-time signals x 1[n] and x 2 [n] shown in Fig. 1-22, represent each of the following
signals by a graph and by a sequence of numbers.
(a) y 1[n] = x 1[n] + x 2 [n]; (b) y2 [n] = 2x 1[n]; (c) y 3 [n] = x 1[n]x 2 [n]
x 1 [n]
3 3
-2 - 1 0 1 2 3 4 5 6 7 n -3 n
Fig. 1-22
y 1[n] = { .. . , 0, - 2, - 2, 3, 4, 3, - 2, 0, 2, 2, 0, .. .}
t
(b) y 2[n] is sketched in Fig . 1-23(b). From Fig. 1-23(b) we obtain
y2 [n] = { .. . , 0, 2, 4, 6, 0, 0, 4, 4, 0, .. .}
t
(c) y 3[n] is sketched in Fig . 1-23(c) . From Fig . 1-23(c) we obtain
y 3 [n] = { . . ., 0, 2, 4, 0, ... }
t
4 4
2 2
- 2- 1 3
1 2 4 5 6 7 n -2- 1 0 1 2 3 4 5 6 7 n
-2
(a) (b)
- 2- 1 0 1 2 3 4 5 6 7 n
(c)
Fig. 1-23
1.5. Sketch and label the even and odd components of the signals shown in Fig. 1-24.
Using Eqs . (1.5) and (1.6), the even and odd components of the signals shown in Fig . 1-24 are sketched in Fig. 1-25.
x(t) x(t)
4 4
0 1 2 3 4 5 0
(a) (b)
x[n] x [n]
4 4
0 1 2 3 4 5 6 n - 1 0 1 2 3 4 5 n
(c) (d)
Fig. 1-24
xe(t) x 0 (t)
4 4
-5 0 5
(a)
4 4
(b)
4 4
2 2
- 5 - 4- 3 - 2- 1
- 5 - 4- 3 - 2- 1 0 1 2 3 4 5 n 2 3 4 5 n
-2
(c)
x0 [n]
4 4
- 4- 3 - 2- 1
- 4- 3 - 2- 1 0 1 2 3 4 n 2 3 4 n
-2
(d)
Fig. 1-25
Let x,(t) and x0 (t) be the even and odd components of ei', respectively.
From Eqs. (1.5) and (1.6) and using Euler's formula, we obtain
1 . .
x (t) = -(e11 +e- Jl) = cost
e 2
x (t) = .!..(e11 - e- 11 ) = 1·sint
0 2
1.7. Show that the product of two even signals or of two odd signals is an even signal and that the product of
an even and an odd signal is an odd signal.
Let x(t) = x 1(t)xz(t). If x 1(t) and xz(t) are both even, then
and x(t) is even. If x 1(t) and xz(t) are both odd, then
and x(t) is odd. Note that in the above proof, variable t represents either a continuous or a discrete variable.
1.8. Show that
(a) If x(t) and x[n] are even, then
k k
,L x[n] = x[O] + 2 ,L x[n] (l.75b)
n=-k n=I
Hence,
a x(t) dt
I -a = ra x(t) dt + ra x(t) dt = 2 ra x(t) dt
J0 J0 J0
Similarly,
k -I k
~ x[n]= ~ x[n]+x[O] + ~x[n]
n=-k n=-k n=I
Hence,
k k k k
~ x[n]= ~x[n]+x[O]+ ~x[n]=x[0]+2 ~x[n]
n=-k n=I n=I n=I
(b) Since x(t) and x[n] are odd, that is, x (-t) = -x(t) and x[ -n] = -x[n], we have
Similarly,
I a x(t)dt=f
-a
0 x(t)dt+ rax(t)dt= rax(-A,)dA,+ rax(t)dt
-a J0 J0 J0
and
k -1 k k k
~ x[n]= ~ x[n]+x[O]+ ~x[n]= ~ x[-m]+x[O]+ ~x[n]
n=-k n=-k n=I m=I n=I
k k k k
=- ~ x[m]+x[O]+ ~x[n]=- ~x[n]+x[O]+ ~x[n]
m=I n=I n=I n=I
=x[O]=O
x(t) = ejw,t
Since
we must have
(1.78)
If%= 0, then x(t) = 1, which is periodic for any value of T. If% i= 0, Eq. (l.78) holds if
2:rc
or T=m- m = positive integer
wo
Thus, the fundamental period T0 , the smallest positive T, of x(t) is given by 2:rc/%·
x(t) = cos(w 0 t + 8)
COS[%(t + T) + 8] = cos(%t + 8)
We note that
if
2n
or T=m- m = positive integer
wo
or
(1.79)
or
. al numb ers
-Qo = -m = ration
2n N (l.80)
with radian frequency% and fundamental period T0 = 2Ttl<flo. Consider the discrete-time sequence x[n]
obtained by uniform sampling of x(t) with sampling interval T,. That is,
m = positive integer
or
T, m " anum
-=-=ration l ber (1.81)
To No
Thus, x[n] is periodic if the ratio T,IT0 of the sampling interval and the fundamental period of x(t) is a rational
number.
Note that the above condition is also true for sinusoidal signals x(t) = cos(o1af + (}).
(a) Find the value of sampling interval T, such that x[n] = x(nT,) is a periodic sequence.
(b) Find the fundamental period of x[n] = x(nT,) if T, = 0.1.ir seconds.
(a) The fundamental period of x(t) is T0 = 2.ir/% = 2.ir/15. By Eq. (l.81),x[n] = x(nT,) is periodic if
T,=~=.!!!... (l.82)
T0 2.ir/15 N0
where m and N0 are positive integers. Thus, the required value of T, is given by
m m 2n
T,=-T0 = - - (1.83)
N0 N 0 15
T, = n/10 =~=~
T0 2.ir/15 20 4
0 T 4
N 0 =m-=m-
T5 3
The smallest positive integer N0 is obtained with m = 3. Thus, the fundamental period of x[n] = x(O.lnn)
isN0 = 4.
1.14. Let x 1(t) and x2(t) be periodic signals with fundamental periods T1 and T2 , respectively. Under what
conditions is the sum x(t) = x 1(t) + x 2(t) periodic, and what is the fundamental period of x(t) if it is
periodic?
Since x 1(t) and xz(t) are periodic with fundamental periods T1 and T2 , respectively, we have
Thus,
(1.84)
or
T, k . al numb er
ration
- = - = (1.85)
T2 m
In other words, the sum of two periodic signals is periodic only if the ratio of their respective periods can be
expressed as a rational number. Then the fundamental period is the least common multiple of T1 and T2 , and it is
given by Eq. (1.84) if the integers m and k are relative prime. If the ratio T/T2 is an irrational number, then the
signals x 1(t) and x 2(t) do not have a common period and x(t) cannot be periodic.
1.15. Let x 1[n] and x 2 [n] be periodic sequences with fundamental periods N 1 and N2 , respectively. Under what
conditions is the sum x[n] = x 1[n] + x 2 [n] periodic, and what is the fundamental period of x[n] if it is
periodic?
Since x 1[n] and x 2 [n] are periodic with fundamental periods N 1 and N 2, respectively, we have
Thus,
(l.86)
Since we can always find integers m and k to satisfy Eq. ( 1.86), it follows that the sum of two periodic sequences is
also periodic and its fundamental period is the least common multiple of N 1 and N2 •
1.16. Determine whether or not each of the following signals is periodic. If a signal is periodic, determine its
fundamental period.
1
(g) x[n] = ej(Hl4)n (h) x[n] = cos-n
4
(i) x [n] = cos-n
Jl
+ sm-n
.Jl
(j) x[n] = cos 2 !!_n
3 4 8
where x 1(t) = cos(n/3)t = cos w/ is periodic with T1 = 2n/w 1 = 6 and xp) = sin(n/4)t = sin Wl is periodic
with T2 = 2n/w2 = 8. Since T1 /T2 = ~ = %is a rational number, x(t) is periodic with fundamental period
T0 = 4T1 = 3T2 = 24.
(d) x(t) =cost+ sin [it= x 1(t) + xz(t)
where x 1(t) =cost= cos w 1t is periodic with T1 = 2n/w 1 = 2nand x2(t) = sin flt= sin w2t is periodic
with T2 = 2n/w2 = f l n. Since T1 /T2 = f l is an irrational number, x(t) is nonperiodic.
(e) Using the trigonometric identity sin2 (} = to - cos 2(}), we can write
x(t) = sin 2 t = .!. - .!.cos 2t = x 1(t) + x 2 (t)
2 2
t
where x 1(t) = is a de signal with an arbitrary period and xz(t) = - cos 2t = - t t cos w 2t is periodic with
T2 = 2n/w2 = n. Thus, x(t) is periodic with fundamental period T0 = n.
(f) x(t) = ej[(ir/2)1-11 = e-jej(Jr/2)1 = e-jejw,1--+ wo = ~
Since Q0 /2n =t is a rational number, x[n] is periodic, and by Eq. (1.55) the fundamental period is N 0 = 8.
(h) x[n] = cos tn = cos Q n--+ Q = t0 0
where
1r 1r
x 1[n] = cos-n =cos Q 1n--+ Q 1 = -
3 3
x 2 [n] = sin~n =cos Q 2 n--+ Q 2 = ~
4 4
t
Since Q/2n = (= rational number),x 1[n] is periodic with fundamental period N 1 = 6, and since Qz12n = t
(=rational number),x2 [n] is periodic with fundamental periodN2 = 8. Thus, from the result of Prob. 1.15,x[n]
is periodic and its fundamental period is given by the least common multiple of 6 and 8, that is, N0 = 24.
(j) Using the trigonometric identity cos 2 (} = to + cos 2(}), we can write
2:n: 1 1 :n:
x[n] =cos -n = - +-cos -n = x 1[n] + x 2 [n]
8 2 2 4
where x 1[n] = = t t<l)"is periodic with fundamental period N 1 = 1 and x2 [n] = cos(n/4)n = cos t t
t
Q 2n--+ Q 2 = n/4. Since Q 2 /2n = (=rational number), x2 [n] is periodic with fundamental period N2 = 8.
Thus, x[n] is periodic with fundamental period N 0 = 8 (the least common multiple of N 1 and N2).
JaPx(t) dt = Ja+T
13 x(t) dt +T (l.87)
fT fa+T (l.88)
J0 x(t) dt = a x(t) dt
and
fi Ifi+T Ifi+T Ifi+T
I x(t) dt = x(r - T) dr = x( r) dr = x(t) dt
a a+T a+T a+T
Next, the right-hand side of Eq. (l.88) can be written as
I aox(t) dt = IT
a+T
x(t) dt
Thus,
I a+T x(t) dt =
IT x(t) dt + ia+T x(t) dt
a a+T 0
= i 0a+T x(t) dt + IT
a+T
x(t) dt = iT x(t) dt
0
1.18. Show that if x(t) is periodic with fundamental period T0 , then the normalized average power P of x(t)
defined by Eq. (1.15) is the same as the average power of x(t) over any interval of length T0 , that is,
1 Ji
P=- I
rTo x(t) 12 dt (l.89)
To o
By Eq. (l.15)
. 1 IT/2 2
P= hm - lx(t)I dt
T-oo T -T/2
Allowing the limit to be taken in a manner such that Tis an integral multiple of the fundamental period,
T = kT0 , the total normalized energy content of x(t) over an interval of length Tis k times the normalized energy
content over one period. Then
I I
rTo x(t) 12 dt ] = 1- J,rTo x(t) 12 dt
. [ -1k J,
P= hm
k-oo kTo o To o
1.19. The following equalities are used on many occasions in this text. Prove their validity.
(a)
N-1,Lan= r-aN
1-a
a;tl
(1.90)
n=O N a=l
"'
(b) ,Lan =-1- lal<l (l.91)
n=O 1-a
00 k
(c) ,Lan=~ lal<l (l.92)
n=k 1-a
00
(a) Let
N-1
~
S =~a=
n 1 +a+a2 +···+aN-1
(1.94)
n=O
Then
N-1
aS = a }: an = a+ a 2 + a 3 + ··· + aN (1.95)
n=O
Subtracting Eq. (l.95) from Eq. (l.94), we obtain
(1 - a) S = 1 - aN
Hence if a * 1, we have
N-1 N
S= ~an=l-a (l.96)
n=O 1-a
If a= 1, then by Eq. (l.94)
N-1
~an =l+l+l+···+l=N
n=O
"' n . N-I n . 1- aN
~a= hm ~a= hm - - = - -
n=O N-oo n=O N-oo 1-a 1-a
"'
~an= ak + ak+I + ak+2 + ...
n=k
"' k
k ~ a
=a k(l +a+a 2 +···=a
) ~a
n
=--
n=O 1-a
( d) Taking the derivative of both sides of Eq. ( 1.91) with respect to a, we have
d (f nl
da n=O a =
d ( 1 ) 1
da 1- a = (1- a) 2
and
Hence,
1 "' n 1 ~ a
~ n=O
~ na
n
or ~ na (l-a)2
= (1 - a)2 =
n=O
1.20. Determine whether the following signals are energy signals, power signals, or neither.
(a) x(t) = e- 01u(t), a> 0 (b) x(t) =A cos(aV + (})
(c) x(t) = tu(t) (d) x[n] = (-0.5)nu[n]
(e) x[n] = u[n] (f) x[n] = 2ej 3n
Thus, x(t) is a power signal. Note that periodic signals are, in general, power signals.
(c) E = Jim
T-oo
J T/2
- T/2
2
lx(t)I dt = Jim
T-oo
JT/2 2
0
(T /2) 3
t dt = Jim - - = oo
T- oo 3
.
P = hm -
1 JT/2 2 .
lx(t)I dt = hm -
1 JT/2 t 2 . 1 (T /2) 3
dt = hm - - - = oo
T- oo T - T/2 T-oo T 0 T-oo T 3
Basic Slgnals
u(-t)={~
t>O
(1.97)
t<O
Since T > 0 and T < 0 imply, respectively, that t < 0 and t > 0, we obtain
u(- t) = {~
t>O
t<O
which is shown in Fig . 1-26.
u(-t)
Fig. 1-26
1.22. A continuous-time signal x( t) is shown in Fig. 1-27. Sketch and label each of the following signals.
x(t)
-1 0 2
Fig. 1-27
u(l - t) = {~ t<1
t>1
{~
O < t~l
u(t) - u(t - 1) =
otherwise
- 1 0 0
(a) (b)
x(t)8(t - 3/2)
2
0 2
(c)
Fig. 1-28
1.23. A discrete-time signal x[n] is shown in Fig. 1-29. Sketch and label each of the following signals.
- 4- 3- 2- 1 O 1 2 3 4 5 n
Fig. 1-29
{~
n~l
u[l - n] =
n>l
{~
- 2~n<O
u[n + 2] - u[n] =
otherwise
3 3
- 4 - 3- 2 - 1 0 1 2 3 n -3-2- 1 0 1 2 3 n
(a) (b)
x[n] S[n - 1]
- 2- 1 O 1 2 3 4 n
(c)
Fig. 1-30
1.24. The unit step function u(t) can be defined as a generalized function by the following relation:
u(t) = {~ t<O
Rewriting Eq. (l.98) as
we obtain
f 00
i/J(t)u(t) dt = J: l/J(t)[l - u(t)] dt
f 00
i/J(t)u(t) dt = 0 and J; l/J(t)[l - u(t)] dt = 0
that is,
{~
t>O
u(t)=
t<O
(l.99)
If a> 0,
If a< 0,
J oo
- 00
i/J(t)()(at) dt = -1
a
J-oo ip (-T
00 a
) ()(-r) d-r = - -1
a
Joo ip (-T
-oo a
) ()(-r) d-r
= - .!..ip(~)I = _1 l/J(O)
a a -r=O Ial
Thus, for any a
Joo ijJ(t)()(at) dt
_ 00
1
= ~l/J(O)
Now, using Eq. (1.20) for </J(O), we obtain
J oo _ 00
1
</J(t)()(at) dt = ~</J(O) = ~
1 Joo_
00 </J(t)()(t) dt
= J oo </J(t)~(j(t)
-00
1
dt
if x(t) is continuous at t = t0 .
(b) Verify Eq. (l.25):
if x(t) is continuous at t = 0.
= J~00 </J(t)[x(to)(j(t-to)l dt
for all </J(t) which are continuous at t = t 0 • Hence, by the equivalence property (l.99) we conclude that
x(t)()(t) = x(O)()(t)
du(t)
<5(t) = u'(t) = - -
dt
where l/J(t) is a testing function which is continuous at t = 0 and vanishes outside some fixed interval. Thus, ijJ'(t)
exists and is integrable over 0 < t < oo and ijJ(oo) = 0. Then using Eq. (l.98) or definition (l.18), we have
1.29. Show that the following properties hold for the derivative of <5(t):
(a) J: </J(t)i5'(t) dt
00
= - </J'(O) where </J'(O) = d</J(t)
dt
I
t=O
(l.101)
t6'(t) = -(j(t)
(a) f 1(3t
2 +1)<5(t)dt
(b) J, (3t
2 2 + 1)<5(t) dt
(d) J: e 00
-t <5(2t - 2) dt
(e) J: e00
-t <5'(t) dt
(a) By Eq. (l.21), with a = -1 and b = 1, we have
= J"' e- __!_c5(t
-oo
1
121
-1) dt = .!.e-
2
11
t=I 2e
I "' e- c5'(t)dt=-.!!._(e-
-oo
1
dt
)1 1
t=O
=e-tl
t=O
=l
1.31. Find and sketch the first derivatives of the following signals:
(a) x(t) = u(t) - u(t - a), a> 0
(b) x(t) = t[u(t) - u(t - a)], a >0
t>O
(c) x(t) = sgn t = { 1
-1 t<O
Then
x'(t) = u'(t) - u'(t - a) = c5(t) - c5(t - a)
Thus,
x'(t) = u(t) - u(t - a) - ac5 (t - a)
0 a 0 a 0
--------l - 1
x' (t)
8(t)
a
x' (t)
t x' (t)
2' (t)
0 O a 0
- 8 (t - a)
- a 8 (t - a)
(a) (b) (c)
Fig. 1-31
1.32. Consider the RC circuit shown in Fig. 1-32. Find the relationship between the inputx(t) and the output y(t)
(a) If x(t) = v,(t) and y(t) = v/t).
(b) If x(t) = v,(t) and y(t) = i(t).
+T
v.(t)
~ c
-1
vc(t)
(a) Applying Kirchhoff's voltage law to the RC circuit in Fig . 1-32, we obtain
Letting v.(t) = x (t) and vp) = y (t) and substituting Eq. (1 .04) into Eq. (1.103), we obtain
or
dy (t) 1 1
- - + -y(t) = - x (t) (1.105)
dt RC RC
Thus, the input-output relationship of the RC circuit is described by a first-order linear differential equation
with constant coefficients.
(b) Integrating Eq. (1.104), we have
vc<t) = -1
c J'- co 1(r)dr
. (1.106)
Substituting Eq. (1.106) into Eq. (1.103) and letting v.(t) = x(t) and i(t) = y(t), we obtain
Ry(t)+-
1
c J'- co y(r)dr = x(t)
or
1
y(t)+-
RC - co
J' 1
y(r)dr = -x(t)
R
Differentiating both sides of the above equation with respect tot, we obtain
Thus, the input-output relationship is described by another first-order linear differential equation with
constant coefficients.
1.33. Consider the capacitor shown in Fig. 1-33. Let input x(t) = i(t) and output y(t) = v/t).
(a) Find the input-output relationship.
(b) Determine whether the system is (i) memoryless, (iz) causal, (iiz) linear, (iv) time-invariant, or (v) stable.
i(t)
Fig. 1-33
(a) Assume the capacitance C is constant. The output voltage y(t) across the capacitor and the input current x(t)
are related by [Eq. (1.106)]
y(t) = T{x(t)} = -1
c J'-co x(r)dr (1.108)
(b) (i) From Eq. (1.108) it is seen that the output y(t) depends on the past and the present values of the input.
Thus, the system is not memoryless.
(ii) Since the output y(t) does not depend on the future values of the input, the system is causal.
(iii) Let x(t) = a 1x 1(t) + a 2 x 2(t). Then
= a 1y 1(t) + a 2 Yz(t)
Thus, the superposition property (1.68) is satisfied and the system is linear.
(iv) Let y 1(t) be the output produced by the shifted input current x 1(t) = x(t - t0).
Then
y 1(t) = T{x(t - t0 )} = -1
c f'
-00
x(r - t0 )dr
(1.109)
where r(t) = tu(t) is known as the unit ramp function (Fig. 1-34). Since y(t) grows linearly in time
without bound, the system is not BIBO stable .
r(t) = tu(t)
0
Fig. 1-34 Unit ramp function .
1.34. Consider the system shown in Fig. 1-35. Determine whether it is (a) memoryless, (b) causal, (c) linear,
(d) time-invariant, or (e) stable.
Multiplier
x(t) y(f) = X(f) COS WC f
Fig. 1-35
Since the value of the output y(t) depends on only the present values of the input x(t), the system is memoryless.
(b) Since the output y(t) does not depend on the future values of the input x(t), the system is causal.
(c) Let x(t) = a 1x(t) + a 2 x(t). Then
y(t) = T{x(t)} = [a 1x 1(t) + a 2 x2 (t)] cos wet
= a 1x 1(t) cos wet+ a 2x2 (t) cos wet
= a 1y1(t) + a 2 Yz(t)
Thus, the superposition property (1.68) is satisfied and the system is linear.
(d) Lety 1(t) be the output produced by the shifted inputx1(t) = x(t- t0 ). Then
But
Thus, if the input x(t) is bounded, then the output y(t) is also bounded and the system is BIBO stable.
y = T{x} = x 2 (l.110)
1.36. The discrete-time system shown in Fig. 1-36 is known as the unit delay element. Determine whether the
system is (a) memoryless, (b) causal, (c) linear, (d) time-invariant, or (e) stable.
__x[n]
____,.~ I Unit .ly[n] =x[n-1] .."'
. delay
Since the output value at n depends on the input values at n - 1, the system is not memoryless.
(b) Since the output does not depend on the future input values, the system is causal.
(c) Letx[n] = a 1x 1[n] + a 2 x2[n]. Then
y[n] = T{a 1x 1[n] + a zX2[n]} = a 1x 1[n - 1] + azX2[n - 1]
= a1Y1[n]+a2Yz[n]
Thus, the superposition property (1.68) is satisfied and the system is linear.
(d) Let y 1[n] be the response to x 1[n] = x[n - n0]. Then
and
Hence, the system is time-invariant.
(e) Since
+
x[n] Unit y[n]
delay
Fig. 1-37
From Fig. 1-37 the input to the unit delay element is x[n] - y[n]. Thus , the output y[n] of the unit delay element is
[Eq. (1.111)]
y[n] = x[ n - 1] - y[n - 1]
Rearranging, we obtain
Thus, the input-output relation of the system is described by a first-order difference equation with constant
coefficients.
(a) Since the output value at n depends on only the input value at n, the system is memoryless.
(b) Since the output does not depend on the future input values, the system is causal.
(c) Letx[n] = a 1x 1[n] + a 2x2[n]. Then
y[n] = T{x[n]} = n{a 1x 1[n] + a 2 x 2 [n]}
= a 1nx 1[n] + a 2nx2[n] = a 1y 1[n] + a 2y2[n]
Thus, the superposition property (1.68) is satisfied and the system is linear.
(d) Lety 1[n] be the response tox 1[n] = x[n - n0 ]. Then
- 2 - 1 0 1 2 3 4 n - 2 - 1 0 1 2 3 4 n
Fig. 1-38
1.39. A system has the input-output relation given by
Hence , the system is not time-invariant unless k0 = 1. Note that the system described by Eq. (1.114) is called a
compressor. It creates the output sequence by selecting every k0th sample of the input sequence . Thus , it is obvious
that this system is time-varying.
1.40. Consider the system whose input-output relation is given by the linear equation
y=ax+b (1.115)
where x and y are the input and output of the system, respectively, and a and bare constants. Is this
system linear?
If b =fa 0, then the system is not linear because x = 0 implies y = b =fa 0. If b = 0 , then the system is linear.
1.41. The system represented by Tin Fig. 1-39 is known to be time-invariant. When the inputs to the system
are x 1[n], x 2 [n], and x 3 [n], the outputs of the system are y 1[n], y 2 [n], and y 3 [n] as shown. Determine
whether the system is linear.
y, [n]
2 2
-2 - 1 0 1 2 3 4 n -2- 1 0 1 2 3 4 n
---+0--+ 2
-2 - 1 0 2 3 4 n -2- 1 0 1 2 3 4 n
x3(nj Y3[n]
3
2
---+0--+
-2 - 1 0 1 2 3 4 n -2- 1 0 1 2 3 4 n
Fig. 1-39
From Fig. 1-39 it is seen that
which is shown in Fig. 1-40. From Figs. 1-39 and 1-40 we see that
2 2 2
+
-2- 1 0 1 2 3 4 n - 2- 1 0 1 2 3 4 n -2- 1 0 1 2 3 4 n
Fig. 1-40
1.42. Give an example of a system that satisfies the condition of additivity ( 1.66) but not the condition of
homogeneity (1.67).
1.43. (a) Show that the causality for a continuous-time linear system is equivalent to the following statement:
For any time t0 and any input x(t) with x(t) = 0 fort :5 t0 , the output y(t) is zero fort :5 t0 .
(b) Find a nonlinear system that is causal but does not satisfy this condition.
(c) Find a nonlinear system that satisfies this condition but is not causal.
(a) Since the system is linear, if x(t) = 0 for all t, then y(t) = 0 for all t. Thus, if the system is causal, then
x(t)= 0 fort,;:; t0 implies that y(t) = 0 fort,;:; t0 • This is the necessary condition. That this condition is
also sufficient is shown as follows: let x 1(t) and x 2(t) be two inputs of the system and let y 1(t) and y 2(t) be
the corresponding outputs. If x 1(t) = x 2(t) fort~ t0 , or x(t) = x 1(t) - xz(t) = 0 fort~ t0 , theny 1(t)= yz(t)
fort~ t 0 , ory(t) = y 1(t) - yz(t) = 0 fort~ t 0 •
y(t) = x(t) + 1
This system is nonlinear (Prob. 1.40) and causal since the value of y (t) depends on only the present value of
= 0 fort~ t 0 , y(t) = 1 fort~ t 0 •
x(t). But with x(t)
y (t) = x(t)x(t + 1)
It is obvious that this system is nonlinear (see Prob. 1.35) and noncausal since the value of y(t) at time t depends
on the value of x(t + 1) of the input at time t + 1. Yet x(t) = 0 fort~ t0 implies that y(t) = 0 fort~ t0 •
(l.117)
Let y(t) be the output of the system with input x(t) = est. Then
T{est} = y(t)
Setting t = 0, we obtain
(l.118)
or T{est} = A est
where/... = y(O).
(1.119)
T{z"} = y[n]
is called an eigenfunction (or characteristic function) of the operator T , and the constant /... is called the eigenvalue
(or characteristic value) corresponding to the eigenfunction x(-). Thus , Eqs. (1 .117) and (1.119) indicate that the
complex exponential functions are eigenfunctions of any LTI system.