Sunteți pe pagina 1din 4

t=[0:dt:20]; % Time axis.

v=rand(size(t))<0.2; % Impulse process.


h=t.*exp(-t.∧ 2); % Pulse shape.
x=conv(v,h); % Shot noise process.
plot(t,x(1:length(t))) % Plot results.
EEM 409 t’); ylabel(‘X(t)’)
xlabel(‘time,

Random Signals

Problem Set-1: (Random Processes and Correlation Functions,


Stationarity and Ergodicity)

Exercises
Problem – 1:
8.1 A random process X(t) consists of three-member functions: x1 (t) = 1,
x2 (t) = −3, and x3 (t) = sin(2π t). Each member function occurs with equal
probability.
(a) Find the mean function, µX (t).
(b) Find the autocorrelation function, RX,X (t1 , t2 ).
(c) Is the process WSS? Is it stationary in the strict sense?
Chapter 8 Random Processes

Problem – 2: Chapter 8 Random Processes

8.2 A random process X(t) has the following member functions: x1 (t) =
−2 cos(t), x (t) = −2sin(t), x3 (t) = 2[cos(t)+sin(t)], x4 (t) = [cos(t)−sin(t)],
8.2 A random2 process X(t) has the following member functions: x1 (t) =
and x5 (t) = [sin(t) − cos(t)].
−2 cos(t), x2 (t) = −2sin(t), x3 (t) = 2[cos(t)+sin(t)], x4 (t) = [cos(t)−sin(t)],
and x5 (t)the
(a) Find = [sin(t) − cos(t)]. µX (t).
mean function,
(b)
(a) Find the
Find the autocorrelation µX (t). RX,X (t1 , t2 ).
mean function, function,
(c)
(b) Is the the
Find process WSS? Is it stationary
autocorrelation function, in
R the(tstrict
, t ). sense?
X,X 1 2
(c) Is
8.3 Problem
Let the –process
a discrete WSS?process
3: random Is it stationary in the strictbysense?
X[n] be generated repeated tosses of a
fair die. Let the values of the random process be equal to the results of each
Let a discrete random process X[n] be generated by repeated tosses of a
8.3 toss.
fair die. Let the values of the random process be equal to the results of each
(a) Find the mean function, µX [n].
toss.
(b) Find the autocorrelation function, RX,X (k1 , k2 ).
(a) Find the mean function, µX [n].
(c) Is the process WSS? Is it stationary in the strict sense?
(b) Find the autocorrelation function, RX,X (k1 , k2 ).
8.4 A (c)discrete
Is the process
randomWSS? Is itX[n],
process, stationary in the by
is generated strict sense?tosses of a coin.
repeated
Let the occurrence of a head be denoted by 1 and that of a tail by −1. A new
8.4  discrete
A discrete random
random process,
process X[n], is generated
is generated by Y[2n] = byX[n]
repeated
for n =tosses
0, ±1,of±2,
a coin.
...
Let the occurrence of a head be denoted by 1 and that of a tail by
by and Y[n] = X[n + 1] for ±n for n odd. Find the autocorrelation function −1. A new
Y[n].random process is generated by Y[2n] = X[n] for n = 0, ±1, ±2, . . .
discrete
for
by and Y[n] = X[n + 1] for ±n for n odd. Find the autocorrelation function
for Y[n].
8.5 Let X[n] be a wide sense stationary, discrete random process with
autocorrelation function RXX [n], and let c be a constant.
8.5 Let X[n] be a wide sense stationary, discrete random process with
(a) Find the autocorrelation function for the discrete random process
autocorrelation function RXX [n], and let c be a constant.
Y[n] = X[n] + c.
(c) Is the process WSS? Is it stationary in the strict sense?
8.4 A (b)discrete
Find the autocorrelation
random process, X[n], function, RX,X (kby
is generated 1, k2 ).
repeated tosses of a coin.
(c) the
Is the process of WSS? Is itbestationary in1the
andstrict sense?
8.4 Let
ExercisesA discrete occurrence a head
random process, denoted
X[n], by
is generated that of a tail by
by repeated −1. Aofnew
tosses a coin. 31
discrete random process is generated by Y[2n] = X[n] for n = 0, ±1, ±2, . . .
Let the occurrence of a head be denoted by 1 and that of a tail by −1. A new
8.4 by A discrete
and Y[n]random 1] for ±nX[n],
= X[n +process, for nisodd.
generated
Find theby repeated tosses
autocorrelation of a coin.
function
discrete random process is generated by Y[2n] = X[n] for n = 0, ±1, ±2, . . .
LetY[n].
for
Problem the occurrence
– 4:= of a head be denoted by 1 and that of a tail by −1. A new
byY(t) areY[n]
and not. Hint:
X[nLet + 1]A(t)for and
±n forB(t)n be independent,
odd. wide sense stationary
Find the autocorrelation function
discrete random process is generated by Y[2n] = X[n] for n = 0, ±1, ±2, . . .
random
for Y[n]. processes with zero means and identical autocorrelation functions.
8.5 Let by andX[n]
Y[n] be =a X[n
wide + 1] sense
for ±n stationary,
for n odd.discrete
Find therandom process function
autocorrelation with
Then let X(t) = A(t) sin(t) and Y(t) = B(t) cos(t). Show that X(t) and Y(t)
autocorrelation
for Y[n]. function RXX [n], and let c be a constant.
8.5 Let are not
X[n]wide be sense
a wide stationary. Then show that
sense stationary, Z(t) israndom
discrete wide sense stationary.
process with
(a) Find the autocorrelation function
autocorrelation function RXX [n], and let c be a constant.for the discrete random process
8.5 Let Y[n] X[n] be a widec. 0 tsense stationary, discrete random process with
8.9 Let X(t) = = X[n]
A(t) +cos(ω + θ ), where A(t) is a wide sense stationary random
autocorrelation
(a) FindX[n]theand function RXX [n],
autocorrelation c bethe
and let for a constant.
(b)process
Are independent of θ , andfunction
Y[n] independent? let θUncorrelated?
be a random discrete
variable
Orthogonal? random process
distributed uni-
Y[n] = X[n] + c.
formly over [0, 2π ). Define a related process Y(t) = A(t) cos(ω0 + ωprocess  
1 )t + θ ).
 (a) Find the autocorrelation function for the discrete random
8.6 (b)
A wide
Are X[n]
sense +Y[n]
stationary, discrete random in process X[n] hasbutan that
autocorre-
Show
Problem that
Y[n] 5:and
=–X(t)
X[n] and c.Y(t)independent?
are stationary Uncorrelated?
the wide Orthogonal?
sense the cross-
lation function
correlation
(b) Are X[n] XYof(t,R
Rand +[k].
tXX
Y[n] ), Find
between
τindependent?the expected
X(t) and value
Y(t), is
Uncorrelated?
Y[n]a function
ofnot = (X[n + m] −
Orthogonal? of τ only
2
8.6 X[n
A and, m])
−hence,
wide sense, where =mX(t)
stationary,
Z(t) is an+arbitrary
discrete
Y(t) is not integer.
random process
stationary in theX[n]
widehassense.
an autocorre-
8.6 lation
A widefunction of RXX [k].discrete
sense stationary, Find therandom
expected valueX[n]
process of Y[n] an(X[n
has = + m] −
autocorre-
8.7
8.10 A random
Let X(t)
X[n m]) 2process
be, awhere is
modifiedgiven by
is version X(t) = A random telegraph process. TheBpro-
cos(ωt)
of theinteger. + B sin(ωt), where A and
lation− function of Rm an arbitrary
XX [k]. Find the expected value of Y[n] = (X[n + m] −
are
cessindependent
switches zero mean
between
2 , where therandom variables.
two states X(t) = 1 and X(t) = −1, with  
X[n −
Problem m]) – 6: m is an arbitrary integer.
therandom
8.7 A
(a) timethe
Find between
process switches
is given µ
mean function, byfollowing
X (t).
exponential
X(t) = A cos(ωt) distributions
+ B sin(ωt), where AfT and
(s) =
B
λ exp(−λ)u(s).
8.7 are independent Also,
zero the
mean starting
X(t)state
random is determined
=Rvariables.
AX,X by flipping a biased
A
(b)random
Find theprocess is given
autocorrelation by
function, 1 , t2 ).+ B sin(ωt), where A and B
cos(ωt)
(t
coinindependent
are so that Pr(X(0) 1) = prandom
zero=mean and Pr(X(0) = −1) = 1 − p.
variables.
(c) Under what conditions
(a) Find the mean function, µX (t).(on the variances of A and B) is X(t) WSS?
(a)
(a) Find
(b) Pr(X(t)
Find the mean 1) and Pr(X(t)
= function,
autocorrelation X (t).
= −1).
µfunction, RX,X (t1 , t2 ).
8.8 Show
(b) by
Find example
the mean that the
function,random
µ (t). process Z(t) = X(t) + Y(t) may be a
(b) Under
(c) Find the autocorrelation function, R (t1 ,of
X the variances t2 ).
A and B) is X(t)
wide sense what conditions
stationary process(on
even thoughX,Xthe random processes WSS?
X(t) and
(c) Under
(c) Find the autocorrelation
what conditions (on function,
the variances , t2A).and B) is X(t) WSS?
RX,X (t1of
 
(d) Is this
8.8 Problem
Show by – process
example
6: WSS?
that the random process Z(t) = X(t) + Y(t) may be a
8.8 wide
Showsense stationary
by example thatprocess even though
the random processtheZ(t) random
= X(t)processes
+ Y(t) mayX(t)beand
a
Let s(t)
8.11 wide be a periodic square wave as illustrated in the accompanying
sense stationary process even though the random processes X(t) and figure.
Suppose a random process is created according to X(t) = s(t − T), where
T is a random variable uniformly distributed over (0, 1).

(a) Find the probability mass function of X(t).


(b) Find the mean function, µX (t).
(c) Find the autocorrelation function, RX,X (t1 , t2 ).
(d) Is this process WSS?
s(t)
1
... ...
1 2 t
−1

8.12 Let s(t) be a periodic triangle wave as illustrated in the accompanying


figure. Suppose a random process is created according to X(t) = s(t − T),
where T is a random variable uniformly distributed over (0, 1).
(iv) the various members of the ensemble are not synchronized.
(a) Find the mean function, µX (t).
(b) Find the autocorrelation function, RX,X (t1 , t2 ).
(c) Is this process WSS?
Problem – 7:
8.14 A random process is defined by X(t) = exp(−At)u(t), where A is a random
variable with PDF, fA (a).
Exercises
(a) Find the PDF of X(t) in terms of fA (a).
Exercises
(b) If A is an exponential random variable, with fA (a) = e−a u(a), find µX (t)
andisRaX,X
8.17 If X(t) wide t2 ). Isstationary
(t1 ,sense the process WSS? random process, find the cross-
Gaussian
3
correlation between X(t) and X (t) in terms of the autocorrelation function
Exercises
8.15 R
Let X(t)
8.17 Problem
IfXXW(τn).isbea– wide
8: IIDsense
an sequencestationary GaussianGaussian
of zero-mean random process,
randomfind the cross-
variables with
3
correlation2between X(t) and X (t) in terms of the autocorrelation function
variance σW . Define a discrete time random process X[n] = pX[n−1]+Wn ,
8.18 Two ). mean discrete random processes, X[n] and Y[n], are statistically
RXX (τzero
8.17 n
If = 1, 2,
X(t) is 3,
independent.
. . ., where
a wide Let sense
a new
X[0] = W0 Gaussian
stationary
random
and p is a random
constant.
process be Z[n] =process, X[n] +find
Y[n].the cross-
Let the
correlation between X(t) and X 3 (t) in terms of the autocorrelation function
8.18 (a)
TwoFindzerothe
autocorrelation mean mean discrete
functions random
function, for µX[n] processes,
X [n]. and Y[n] be X[n] and Y[n], are statistically
R (τ ). be Z[n]
independent.
XX Let a new random
(b) Find the autocorrelation!function, "|k|process R [n , ! ].=
n "|k|X[n] + Y[n]. Let the
1 X,X 1 21
autocorrelation functions
8.18 Two zero mean discrete =for X[n]processes,
RXX [k]random , RY[n]
and be
YY [k]X[n]
= and Y[n], . are statistically
!2 "|k| !3 "|k|
independent.
8.16 Let X(t) and Y(t) Let abenew tworandom 1 wide
jointly process be Z[n]
sense 1= X[n]
stationary + Y[n]. random
Gaussian Let the
Find RZZ [k]. Plotfunctions R [k] =
all three autocorrelation , R [k] =
functions .
3 (you may want to use
XX YY
autocorrelation
processes with zero means X[n]
forand2 with Y[n]
andautocorrelation
be and crosscorrelation
MATLAB to help).
functions denoted as RXX (τ ! ), R"YY (τ ), RXY (τ ). Determine
! "|k| the crosscorrela-
Find RZZ [k]. Plot all three2 autocorrelation 1 |k| 2 functions 1 (you may want to use
tion
8.19 Problemfunction
Consider a between
discrete RXX X wide
[k]
time =(t) and Y , (t).
sense RYY [k] = random. processes whose
stationary
MATLAB– to 9: help). 2 3
autocorrelation funcion is of the form
8.19 Find RZZ [k].
Consider Plot all time
a discrete threewide
autocorrelation
sense stationary functions random(you processes
may wantwhose
to use
|k|
MATLAB
autocorrelation to help). funcion RXXis[k]
of= thea form, where |a| < 1.

8.19 Assume
Considerthis processtime
a discrete has
RXXzero-mean.
wide a|k| , Isstationary
[k] = sense the process
where ergodic
|a| random
< 1. in the mean?
processes whose
autocorrelation funcion is of the form
8.20 Let X(t) be
Assume thisa process
wide sense has stationary
zero-mean.random processergodic
Is the process that is ergodic in the
in the mean?
 
mean and the autocorrelation.
RXX [k] =However,
a|k| , where X(t) |a| < 1.zero-mean. Let Y(t) =
is not
Let X(t)where
8.20 CX(t), be a wide
C is asense
randomstationary
variable random process of
independent that is ergodic
X(t) and C isinnotthe
Assume
mean and
zero-mean.
Problem this
– the
10: process
thathasY(t)zero-mean.
autocorrelation.
Show is notHowever,
ergodic inX(t)
Is the process
the is notergodic
mean zero-mean.
or in the Let Y(t) =
mean?
the autocorrelation.
CX(t), where C is a random variable independent of X(t) and C is not
8.20 Let X(t)
8.21 Prove bethe
that
zero-mean. aShow
wide
family sense
that of stationary
differential
Y(t) is not ergodicrandom in theprocess
equations mean orthattheis ergodic in the
autocorrelation.
mean and the autocorrelation. However, X(t) is not zero-mean. Let Y(t) =
d
8.21 CX(t),
Prove that
wherethe Cfamily of differential
is a random PX (0;variable equations
t) + λP Xindependent
(0; t) = 0, of X(t) and C is not
dt
zero-mean. Show that Y(t) is not ergodic in the mean or the autocorrelation.
d d  
PX (i; t) + λPdtXP (0;=t)λP
(i;Xt) + XλP 1;t)
(i X−(0; t),= 0,i = 1, 2, 3, . . . ,
8.21 Prove that dtthe family of differential equations
d
leads to the Poisson
PX (i; t) distribution
+ λPdX (i; t) = λPX (i − 1; t), i = 1, 2, 3, . . . ,
dt PX (0; t) + λPX (0; t) = 0,
dt (λt)i −λt
leads to the Poisson distribution PX (i; t) = e .
d i!
PX (i; t) + λPX (i; t) = λPX (i −i 1; t), i = 1, 2, 3, . . . ,
dt (λt) −λt
8.22 Consider a Poisson counting PXprocess
(i; t) = with earrival . rate λ.
leads to the Poisson distribution i!
Chapter
(b) Find the mean function of the shot noise 8 Random Processes
process.
(c) Find the autocorrelation function of the shot noise process.

8.23 Let X(t) be a Poisson counting process with arrival rate λ. Find Pr(N(t) =
8.30 A shot noise process with random amplitudes is defined by
k|N(t + τ –
Problem )= m), where τ > 0 and m ≥ k.
11:
!∞
8.24 Let X(t) be a WSS random X(t)process
= Awith mean
i h(t − Si ), µX and autocorrelation
function RXX (τ ). Consider forming
i=1a new process according to

X(t + tfrom
where the Si are a sequence of points o ) − X(t)
a Poisson process and the Ai
Y(t) = .
are IID random variables that are also independent
to of the Poisson points.
(a)
(a) Find
Find the
the mean
mean function of X(t).
function of Y(t).
(b)
(b) Find
Find the
the autocorrelation
autocorrelation function X(t). Is Y(t) WSS?
of Y(t).
function of
 
 
8.25 Let Xi (t), i = 1, 2, . . . , n be a sequence of independent Poisson counting
MATLAB  Problem  Set:  
processes with arrival rates λi . Show that the sum of all of these Poisson
processes,
MATLAB
Problem – Exercises
1: n
!
X(t) = Xi (t),
8.31 You are given a member function of a random process as y(t) = 10 sin(2πt+
i=1
π/2 where the amplitude is in volts. Quantize the amplitude of y(t) into
21 levelsa Poisson
is itself with theprocess.
intervals ranging
What is thefrom
arrival rate5 of
−10. tothe
10.5 in process?
sum 1-volt steps.
Consider 100 periods of y(t) and let t take on discrete values given by nts
8.26 where ts = 5 msec.
A workstation is usedConstruct
until ita histogram
fails and isoftheny(t).sent out for repair. The
 
 time between failures, or the length of time the workstation functions
until it needs repair, is a random variable T. Assume the times between
 failures, T , T , . . . T , of the workstations available are independent ran-
1 2 n
dom variables that are identically distributed. For t > 0, let the number of
workstations that have failed be N(t).
(a) If the time between failures of each workstation has an exponential
PDF, then what type of process is N(t)?
(b) Assume that you have just purchased 10 new workstations and that
each has a 90-day warranty. If the mean time between failures (MTBF)
is 250 days, what is the probability that at least one workstation will
fail before the end of the warranty period?

8.27 Suppose the arrival of calls at a switchboard is modeled as a Poisson


process with the rate of calls per minute being λa = 0. 1.

(a) What is the probability that the number of calls arriving in a 10-minute
interval is less than 10?
(b) What is the probability that the number of calls arriving in a 10-minute
interval is less than 10 if λa = 10?

S-ar putea să vă placă și