Download as ppt, pdf, or txt
Download as ppt, pdf, or txt
You are on page 1of 61

Chapter 1 Random Process

1.1 Introduction (Physical phenomenon)


Deterministic model : No uncertainty about its timedependent behavior at any instant of time .
Random model :The future value is subject to
chance(probability)
Example: Thermal noise , Random data stream

1.2 Mathematical Definition of a Random Process (RP)


The properties of RP
a. Function of time.
b. Random in the sense that before conducting an
experiment, not possible to define the waveform.
Sample space S
function of time, X(t,s)
mapping
1

S X (t,s )

-T t T

2T:The total observation interval


s j X (t , s j ) x j (t )

x j (t )= sample function

(1.1)
(1.2)

At t = tk, xj (tk) is a random variable (RV).


To simplify the notation , let X(t,s) = X(t)
X(t):Random process, an ensemble of time function
together with a probability rule.

Difference between RV and RP


RV: The outcome is mapped into a number
RP: The outcome is mapped into a function of time
2

Figure 1.1 An ensemble of sample functions:


{x j (t ) | j 1,2, , n}

1.3 Stationary Process


Stationary Process :
The statistical characterization of a process is independent of
the time at which observation of the process is initiated.
Nonstationary Process:
Not a stationary process (unstable phenomenon )
Consider X(t) which is initiated at t = ,
X(t1),X(t2),X(tk) denote the RV obtained at t1,t2,tk
For the RP to be stationary in the strict sense (strictly stationary)
The joint distribution function
FX ( t1 ),..., X ( tk ) ( x 1,.., xk ) FX ( t1 ) ,...,X ( tk ) ( x 1,... xk ) (1.3)
For all time shift , all k, and all possible choice of t1,t2,tk
4

X(t) and Y(t) are jointly strictly stationary if the joint


finite-dimensional distribution of X (t1 ) X (tkand
)

Y (t ' )Y (t ' )
1

are invariant w.r.t. the origin t = 0.

Special cases of Eq.(1.3)


FX ( t ) ( x ) FX ( t ) ( x ) FX ( x )
1.
for all t and (1.4)
2. k = 2 , = -t1

FX ( t1 ), X ( t2 ) ( x1, x ) FX (0), X ( t2 t1 ) ( x , x ) (1.5)


2

which only depends on t2-t1 (time difference)


5

Figure 1.2 Illustrating the probability of a joint event.

Figure 1.3 Illustrating the concept of stationarity in Example 1.1.


7

1.4 Mean, Correlation,and Covariance Function


Let X(t) be a strictly stationary RP
The mean of X(t) is

X (t ) E X (t )

xf X ( t ) ( x ) d x

for all t

(1.6)
(1.7)

fX(t)(x) : the first order pdf.


The autocorrelation function of X(t) is

R X (t1,t2 ) E X (t1 ) X (t2 )

- -

- -

x1 x2 f X ( t1 ) X ( t2 ) ( x1 , x2 )dx1dx2
x1 x2 f X (0) X ( t2 t1 ) ( x1 , x2 )dx1dx2

RX (t2 t1 )

for all t1 and t2

(1.8)

The autocovariance function

C X (t1,t 2) E X (t1 ) X X (t2 ) X


RX (t2 t1 ) X2

(1.10)

Which is of function of time difference (t2-t1).


We can determine CX(t1,t2) if X and RX(t2-t1) are known.
Note that:
1. X and RX(t2-t1) only provide a partial description.
2. If X(t) = X and RX(t1,t2)=RX(t2-t1),
then X(t) is wide-sense stationary (stationary process).
3. The class of strictly stationary processes with finite
second-order moments is a subclass of the class of all
stationary processes.
4. The first- and second-order moments may not exist.
9

Properties of the autocorrelation function


For convenience of notation , we redefine

RX ( ) E X (t ) X (t ) ,

for all t

(1.11)

1. The mean-square value

RX (0) E X 2 (t ) , 0

(1.12)

2. RX ( ) R( )

(1.13)

3. RX ( ) RX (0)

(1.14)
10

Proof of property 3:
E [( X (t ) X (t )) ] 0
2

Consider

E[ X 2 (t )] 2 E[ X (t ) X (t )] E[ X 2 (t )] 0

2 E [ X (t )] 2 RX ( ) 0
2

2 RX (0) 2 RX ( ) 0
RX (0) RX ( ) RX (0)

| RX ( ) | RX (0)
11

The RX() provides the interdependence information


of two random variables obtained from X(t) at times
seconds apart

12

Example 1.2 X (t ) Acos(2f ct )


1
,
f ( ) 2
0,
elsewhere

(1.15)

(1.16)

A2
RX ( ) E X (t ) X (t )
cos(2f ct )
2

(1.17)

13

Appendix 2.1 Fourier Transform

14

We refer to |G(f)| as the magnitude spectrum of the signal g(t),


and refer to arg {G(f)} as its phase spectrum.

15

DIRAC DELTA FUNCTION


Strictly speaking, the theory of the Fourier transform is
applicable only to time functions that satisfy the Dirichlet
conditions. Such functions include energy signals. However,
it would be highly desirable to extend this theory in two
ways:
1. To combine the Fourier series and Fourier transform into a
unified theory, so that the Fourier series may be treated as a
special case of the Fourier transform.
2. To include power signals (i.e., signals for which the average
power is finite) in the list of signals to which we may apply
the Fourier transform.
16

The Dirac delta function or just delta function, denoted by (t ) ,


is defined as having zero amplitude everywhere except at
, where
t 0 it is infinitely large in such a way that it contains unit
area under its curve; that is

(t ) 0,

and

t0

(t ) dt 1

(A2.3)

(A2.4)

g (t ) (t t0 )dt g (t0 )

(A2.5)

g ( ) (t )d g (t )

(A2.6)

17

18

Example 1.3 Random Binary Wave / Pulse


1. The pulses are represented by A volts (mean=0).
2. The first complete pulse starts at td.
1
, 0 td T
fTd (t d ) T
0, elsewhere

3. During (n 1)T t t d nT, the presence of +A


or A is random.
4.When tk ti T , Tk and Ti are not in the same pulse
interval,
X(t ) and X(ti) are

E X (thence,
X (tindependent.
k ) X ( ti ) kE X ( tk ) E
i ) 0
19

Figure 1.6
Sample function of random binary wave.

20

4. When tk ti T , Tk and Ti are not in the same pulse


interval, hence, X(tk) and X(ti) are independent.
E X (tk ) X (ti ) E X (tk ) E X (ti ) 0

21

5. For tk ti T , tk 0 , ti tk
X(tk) and X(ti) occur in the same pulse interval

iff td T- tk -ti
i.e., td -ti T
A2 , td T - tk - ti

E X (tk ) X (ti ) td

0,

E X (tk ) X (ti )

T- tk -ti
0

elsewhere

A 2 f Td (td )dtd

A2

dtd
0
T
tk ti
2
A (1
)
T
T- tk -ti

tk ti T
22

6. Similar reason for any other value of tk


2

A (1 ), T
RX ( )
, where tk -ti
T

0,
T

What is the Fourier Transform of R (?)


X
Reference : A.Papoulis, Probability, Random
Variables and Stochastic Processes,
Mc Graw-Hill Inc.

23

Cross-correlation Function
RXY (t,u ) E X (t )Y (u )

(1.19)

(1.20)
and RYX (t,u ) E Y (t ) X (u )
Note RXY (t , u ) and RYX (t , u ) are not general even
functions.
The correlation matrix is

RX (t , u ) RXY (t , u )
R (t , u )

R
(
t
,
u
)
R
(
t
,
u
)
Y
YX

If X(t) and Y(t) are jointly stationary


RX ( ) RXY ( )
R ( )
(1.21)

RYX ( ) RY ( )
where t u

24

Proof of RXY ( ) RYX ( ) :


RXY ( ) E[ X (t )Y (t )]
Let t ,
RXY ( ) E [ X ( )Y ( )]
E [Y ( ) X ( )]
E [Y (t ) X (t ( )]
RYX ( )

(1.22)
25

Example 1.4 Quadrature-Modulated Process

X 1 (t ) X (t ) cos( 2f c t )
X 2 is(ta) stationary
X (t ) process
sin( 2f
),
where X(t)
andc tis
uniformly
distributed over [0, 2].
R12 () E X 1 (t ) X 2 (t )

E X (t ) X (t ) E cos(2f c t ) sin( 2f c t 2f c )

1
RX ( ) E sin( 2f c t 2f c 2) E sin( 2f c )
2
1
=0
RX ( ) sin(2f c )
2

(1.23)

At 0, sin(2f c ) 0, R12 ( ) 0 ,
X 1 (t ) and X 2 (t ) are orthogonal.

26

1.5 Ergodic Processes


Ensemble averages of X(t) are averages across the
process.
Long-term averages (time averages) are averages along the
process
DC value of
T (random variable)
1 X(t)
x (T )
x (t ) dt
(1.24)

2T T
? 1
If X(t) is stationary,

E x (T )

2T
1

2T
X

E x (t ) dt

X dt
27
(1.25)

x (T ) represents an unbiased estimate of X

The process X(t) is ergodic in the mean, if


a.

lim x (T ) X

b. lim var x (T ) 0
T

The time-averaged autocorrelation function


1 T
Rx ( ,T )
x (t )x (t )dt

2 T
Rx (, T ) is a random variable.

(1.26)

If the following conditions hold, X(t) is ergodic in the


autocorrelation functions
lim Rx ( , T ) RX ( )
T

lim var Rx (,T ) 0

28

1.6 Transmission of a random Process Through a Linear TimeInvariant Filter (System)

(t )impulse
h(response
1 )X (t of1the
) dsystem
where h(t) is Ythe
1
-

Y ( t ) E Y (t )
If E[X(t)] is finite
and system is stable

h ( 1 ) X ( t 1 ) d1

h ( 1 ) E x ( t 1 ) d1

If X(t) is stationary,

h ( 1 ) X ( t 1 ) d1
H(0) :System DC response.
-

(1.27)

Y X h( 1 ) d1 X H (0),
-

(1.28)

(1.29)

29

Consider autocorrelation function of Y(t):


RY (t, ) E Y (t )Y ( )
E

h( 1 )X (t 1 ) d1 h( 2 ) X ( 2 ) d 2

(1.30)

2
E[
X
(t )] is finite and the system is stable,
If

RY (t,)

d1 h( 1 )

If

d 2 h ( 2 ) RX (t 1, 2 )

(1.31)

RX (t 1 , 2 ) RX (t 1 2 ) (stationary)
RY ( )

h( 1 )h ( 2 ) RX ( 1 2 ) d1 d2

(1.32)

Stationary input, Stationary output

RY (0) E Y (t )
2

h( 1 )h( 2 )RX ( 2 1 ) d1 d 2 (1.33)


30

1.7 Power Spectral Density (PSD)


Consider the Fourier transform of g(t),

G( f )
g (t )

g (t ) exp( j 2ft ) dt

G ( f ) exp( j 2ft ) df

Let H(f ) denote the frequency response,

2 -1

h( 1 ) H ( f ) exp( j 2f1 ) df

h( ) R ( ) d d
H
(
f
)
exp(
j
2

f
)
df
1
2
X
2
1
1
2

E Y (t )
2

(1.34)

df H ( f ) d2h( 2 ) RX ( 2 1 ) exp( j 2f1 ) d1

(1.35)

df H ( f ) d2h(2 ) exp( j 2f 2 ) RX ( ) exp( j 2f ) d


*
H ( f ) (complex conjugate response of the filter)

(1.36)

31

E Y (t ) df H ( f )
2

RX ( ) exp( j 2f ) d

(1.37)

H ( f ) : the magnitude response

Define: Power Spectral Density ( Fourier Transform of R( )

S X ( f ) RX ( ) exp( 2f ) d

E Y (t ) H ( f ) S X ( f ) df
2

(1.38)

(1.39)

Recall E Y (t ) h( 1 )RX ( 2 1 ) d1 d 2
(1.33)
- -
Let H ( f ) be the magnitude response of an ideal narrowband filter
1, f f c 1 f
2
|H ( f )|
(1.40)
1
0, f f c 2 f
2

f : Filter Bandwidth
If f

f c and S X ( f ) is continuous,

E Y 2 (t ) 2f S X ( f c ) in W/Hz

32

Properties of The PSD

S X ( f ) RX ( ) exp( j 2f ) d

RX ( ) S X ( ) exp( j 2f ) df

(1.42)
(1.43)

Einstein-Wiener-Khintahine relations:

S X ( f ) RX ( )
S X ( f ) is more useful than RX ( ) !
33

a. S X (0) RX ( ) d

(1.44)

b. E X (t ) S X ( f ) df
2

(1.45)

c. If X (t ) is stationary,

E Y 2 (t ) (2f ) S X ( f ) 0
SX ( f ) 0

for all f

(1.46)

d. S X ( f ) RX ( ) exp( j 2f ) d

RX (u ) exp( j 2fu ) du,

SX ( f )

u
(1.47)

e. The PSD can be associated with a pdf :

pX ( f )

SX ( f )
S X ( f ) df

(1.48)
34

Example 1.5 Sinusoidal Wave with Random Phase

X (t ) A cos(2fct ), ~ U ( , )
A2
RX ( )
cos(2fc )
2

S X ( f ) RX ( ) exp( j 2f ) d

A2
exp( j 2f c )d exp( j 2f c ) exp( j 2f ) d

4
A2
( f f c ) ( f f c )

4
Appendix 2,

exp j 2 ( f c f ) d ( f f c )

35

Example 1.6 Random Binary Wave (Example 1.3)


if m(t) 1
if m(t) 0

A,
X (t )
A,

A2 (1 )
RX ( )
T

T
T

) exp( j 2f ) d
T
A 2T sinc 2 ( f T )

SX ( f )

A2 (1

(1.50)

Define the energy spectral density of a pulse as


g ( f ) A2T 2 sinc2 ( f T )

(1.51)

g ( f )
T

(1.52)

SX ( f )

36

Example 1.7 Mixing of a Random Process with a Sinusoidal Process


Y (t ) X (t ) cos(2f ct ) , ~ U (0,2 )
RY ( ) E Y (t )Y (t )

(1.53)

E X (t ) X (t ) E cos(2f ct 2f c ) cos(2fct )
1
RX ( ) E cos(2f c ) cos(4f c t 2f c 2)
2
1
RX ( ) cos(2f c )
2

(1.54)

SY ( f ) RY ( ) exp( j 2f ) d

1
RX ( ) exp( j 2 ( f f c )) exp( j 2 ( f f c )) d
4
1
S X ( f f c ) S X ( f f c )
(1.55)
4
We shift the S X ( f )to the right by f c , shift it to the left by f c ,
add them and divide by 4.

37

Relation Among The PSD of The Input and Output Random Processes
X(t)
SX
(f)

h(t)

Y(t)
SY
(f)

Recall (1.32)
RY ( )

SY ( f )

SY ( f )

h( 1 )h( 2 ) RX ( 1 2 ) d 1 d 2

(1.32)

h( 1 )h ( 2 ) RX ( 1 2 ) exp( j 2f ) d 1 d 2 d

Let 1 2 0 , or 0 1 2

h( 1 )h ( 2 ) RX ( 1 ) exp( j 2f 0 ) exp( j 2f 2 ) exp( j 2f 0 ) d1 d 2 d0

SX ( f )H ( f )H * ( f )
2

H ( f ) SX ( f )

(1.58) 38

Relation Among The PSD and


The Magnitude Spectrum of a Sample Function
Let x(t) be a sample function of a stationary and ergodic Process X(t).
In general, the condition for Fourier transformable is

x(t ) dt

(1.59)

This condition can never be


satisfied by any stationary x(t) with infinite duration.
T
(1.60)
We may write X ( f , T ) T x (t ) exp( j 2ft ) dt

Ergodic Take time average


1 T
RX ( ) lim
x (t )x (t ) dt
T 2T T
If x(t) is a power signal (finite average power)
1 T
1
2
x
(
t

)
x
(
t
)
dt

X
(
f
,
T
)
2T T
2T
Time-averaged autocorrelation
periodogram function

(1.61)

(1.62)

39

Take inverse Fourier Transform of right side of (1.62)

1
2

1
2
T x(t )x(t )dt 2T X ( f , T ) exp( j 2 fT )df

(1.63)

From (1.61),(1.63),we have

1
2
R
(

lim
X
(
f
,
T
)
exp( j 2f )df
NoteXthat forT any
2T x(t) periodogram does not converge as
given

(1.64)

Since x(t) is ergodic

1
2
E X ( f T ) exp( j 2f )df
T 2T

1
2
RX ( ) lim
E X ( f T ) exp( j 2f )df
T 2T

E RX ( ) RX ( ) lim

(1.66)

Recall (1.43) RX ( ) S X ( f ) exp( j 2f )df

1
2
S X ( f ) lim
E X ( f ,T )
T 2T
(1.67) is used to estimate the PSD of x(t)
2

1 T
lim
E x (t ) exp( j 2ft )dt
T 2T
T

(1.67)

40

Cross-Spectral Densities

S XY ( f ) RXY ( ) exp( j 2f )d

SYX ( f ) RYX ( ) exp( j 2f )d

(1.68)
(1.69)

S XY ( f ) and SYX ( f ) may not be real.

RXY ( ) S XY ( f ) exp( j 2f )df

RYX ( ) SYX ( f ) exp( j 2f )df

RXY ( ) RYX ( )

S XY ( f ) SYX ( f ) SYX
(f)

(1.22)
(1.72)
41

Example 1.8 X(t) and Y(t) are zero mean stationary processes.
Consider Z (t ) X (t ) Y (t )
S Z ( f ) S X ( f ) SY ( f )
(1.75)
Example 1.9 X(t) and Y(t) are jointly stationary.

RVZ (t , u ) E V (t ) Z (u )
E

h1 ( 1 ) X (t 1 )d 1 h2 ( 2 )Y (u 2 )d 2

h1 ( 1 )h2 ( 2 )RXY (t 1 , u 2 )d 1d 2

Let t u
RVZ ( )

h1 ( 1 )h2 ( 2 )RXY ( 1 2 )d 1d 2 (1.77)

SVY ( f ) H1 ( f ) H 2 ( f ) SXY ( f )

42

1.8 Gaussian Process


Define : Y as a linear functional of X(t)
T
Y g (t ) X (t )dt
( g(t): some function)
0

( e.g g(t):

(1.79)

(e) )

The process X(t) is a Gaussian process if every linear


functional of X(t) is a Gaussian random variable

( y Y ) 2
1
fY ( y )
exp

2
2

2 Y
Y

1
y2
Normalized fY ( y )
exp( ) , as N (0,1)
2
2

(1.80)
(1.81)

Fig. 1.13 Normalized Gaussian distribution

43

Central Limit Theorem


Let Xi , i =1,2,3,.N be (a) statistically
independent R.V.
2
and (b) have mean X and variance X.
Since they are independently and identically distributed (i.i.d.)
Normalized Xi

1
Yi
( X i X )
X

i 1,2,...., N

Hence, EYi 0,

Var Yi 1.

1
Define VN
N

Y
i 1

The Central Limit Theorem


The probability distribution of VN approaches N(0,1)
as N approaches infinity.

44

Properties of A Gaussian Process


1.

X(t)
Gaussian

h(t)

Y(t)
Gaussian

Y (t ) h(t )X ( )d
0

Define

Z gY (t ) h (t ) X ( ) d dt

T
0

gY (t )h(t ) dt X ( ) d

g ( ) X ( ) d
0

where g ( ) gY (t )h (t )dt
0

By definition Z is a Gaussian random variable (1.81)


T

Y (t ) h(t )X ( )d , 0 t is Gaussian
0

45

2. If X(t) is Gaussisan
Then X(t1) , X(t2) , X(t3) , ., X(tn) are jointly Gaussian.
Let X ( ti ) E X (ti ) i 1,2 ,....,n
and the set of covariance functions be

C X (tk , ti ) E X (tk ) X ( tk ) X (ti ) X ( ti ) ,


where X X (t1) ,X (t 2) ,....,X (tn )
Then f X ( t1 ),..., X ( tn ) ( x1 ,..., xn )

k,i 1,2 ,...,n

1
T
1
exp(

(
x

( x )) (1.85)
n
1
2
( 2 ) 2 2

where mean vector 1 , 2 ,...., n

covariance matrix {C X (tk , ti )}nk ,i 1


determinant of covariance matrix

46

3. If a Gaussian process is stationary then it is strictly stationary.


(This follows from Property 2)
4. If X(t1),X(t2),..,X(tn) are uncorrelated as

E [( X (t k ) X ( tk ) )( X (ti ) X ( ti ) )] 0
Then they are independent
Proof : 2uncorrelated
1

2
2
,
where

E
[(
X
(
t
)

E
(
X
(
t
))
] , i 1,2 ,n.
i
i
i

n2

1 is also a diagonal matrix

(1.85) f X ( t1 ),, X ( tn ) ( x1 ,..., xn )

1
n

( 2 )
2

exp(

1
x T 1 (x ))
2

f X ( x ) f X i ( xi )
i 1

xi X i
1

where X i X (ti ) and f X i ( xi )


exp
2

2 i
i

47

1.9 Noise
Shot noise
Thermal noise


1
1
E I
E V 4kT f 4kTGf
R
R
E VTN2 4kTRf
2
TN

2
TN

volts2
amps2

k: Boltzmanns constant = 1.38 x 10-23 joules/K, T is the


absolute temperature in degree Kelvin.
48

White noise

N0
SW ( f )
2
N 0 kTe

(1.93)
(1.94)

Te : equivalent noise temperature of the receiver


N0
RW ( )
( )
2

(1.95)
49

Example 1.10 Ideal Low-Pass Filtered White Noise

N 0
-B f B
SN ( f ) 2
f B
0
B N
0
RN ( )
exp( j 2 f ) df
B 2
N 0 B sinc( 2 B )

(1.96)
(1.97)
50

Example 1.11 Correlation of White Noise with a Sinusoidal Wave


'
w
(t
)
w
(t )
dt
White noise

X
T

2
cos(2fct )
T

fc

2 T
w' ( t )
w( t ) cos( 2f c t )dt

0
T
The varance of w' ( t ) is
2

2 E
T

k
, k is integer
T
(1.98)

w(t1 ) cos(2 f ct1 ) w(t2 ) cos(2 f ct2 ) dt1 dt2

2 T T

E w(t1 ) w(t2 ) cos( 2 f c t1 ) cos( 2 f c t2 ) dt1 dt2


T 0 0
2 T T

RW ( t1 , t2 ) cos( 2 f c t1 ) cos( 2 f c t2 ) dt1 dt2

0
0
T
From (1.95)
2 T T N0
2
(t1 t2 ) cos( 2 f c t1 ) cos( 2 f c t2 ) dt1 dt2
0
0
T
2
N0 T
N0
2
51

cos
(
2

f
t
)
dt

(1.99)
c

0
T
2

1.10 Narrowband Noise (NBN)

Two representations
a. in-phase and quadrature components (cos(2 fct) ,sin(2 fct))
b.envelope and phase
1.11 In-phase and quadrature representation
n (t ) nI (t ) cos(2 f c t ) nQ (t ) sin( 2 f c t )

(1.100)

nI (t ) and nQ (t ) are low - pass signals

52

Important Properties
1.nI(t) and nQ(t) have zero mean.
2.If n(t) is Gaussian then nI(t) and nQ(t) are jointly Gaussian.
3.If n(t) is stationary then nI(t) and nQ(t) are jointly stationary.
SN ( f fc ) SN ( f fc ) ,
4.S N I ( f ) S NQ ( f )
0

5. nI(t) and nQ(t) have the same variance

-B f B
otherwise

(1.101)

N0
2 .

6.Cross-spectral density is purely imaginary.


S N I NQ ( f ) S NQ N I ( f )
j S N f f c S N f f c ,

-B f B
otherwise

(1.102)

7.If n(t) is Gaussian, its PSD is symmetric about fc, then nI(t) and
nQ(t) are statistically independent.
53

Example 1.12 Ideal Band-Pass Filtered White Noise

fc B N
N0
0
RN ( )
exp( j 2 f )df
exp( j 2 f )df
fc B 2
fc B 2
N 0 B sinc( 2 B ) exp( j 2 f c ) exp( j 2 f c )
fc B

2 N 0 B sinc( 2 B ) cos(2 f c )

(1.103)

Compare with (1.97) (a factor of ),


RN I ( ) RN Q ( ) 2 N 0 B sinc( 2 B ).

54

1.12 Representation in Terms of Envelope and Phase Components


n (t ) r (t ) cos 2f c t (t )
(1.105)
Envelope

r (t ) n (t ) n (t )
2
I

2
Q

(1.106)

Phase
nQ (t )
(t ) tan

n
(
t
)
I

(1.107)

Let NI and NQ be R.V.s obtained (at some fixed time) from nI(t)
and nQ(t). NI and NQ are independent Gaussian with zero mean and
variance 2.
55

nI2 nQ2
1
f N I , N Q (nI , nQ )
exp(
)
2
2
2
2

f N I , NQ ( nI , nQ )dnI dnQ
Let nI r cos

nI2 nQ2
1
exp(
) dnI dnQ
2
2
2
2

(1.108)

(1.109)
(1.110)

nQ r sin

(1.111)

dnI dnQ r dr d

(1.112)
56

Substituting (1.110) - (1.112) into (1.109)


f N I , NQ ( nI , nQ )dnI dnQ f R , ( r, ) rdrd
r
r2

exp( 2 ) rdrd
2
2
2
r
r2
f R , ( r, )
exp( 2 )
2
2
2
1
0 2
0 2 , f ( ) 2
elsewhere
0
r
r2

f R ( r ) 2 exp( 2 2 ) , r 0
0
elsewhere
f R ( r ) is Rayleigh distribution.
For convenience , let
fV ( )

(1.113)
(1.114)

(1.115)

r
. fV ( ) f R ( r )

2
exp( ) , 0
2
0
elsewhere

(1.118)

57

Figure 1.22 Normalized Rayleigh


distribution.
58

1.13 Sine Wave Plus Narrowband Noise


x (t ) A cos(2f c t ) n(t )
x (t ) nI (t ) cos(2f c t ) nQ (t ) sin( 2f c t )
nI (t ) A nI (t )

(1.119)

If n(t) is Gaussian with zero mean and variance


1. nI ' (t ) and nQ (t ) are Gaussian and statistically independent.
2.The mean of nI ' (t ) is A and that of nQ (t ) is zero.
2
3.The variance of nI ' (t ) and nQ (t ) is .2
2

( nI A) nQ
1
f N I , N Q (nI , nQ )
exp

2
2
2
2

Let

r (t ) nI (t ) nQ2 (t )
2

1
2

nQ (t )

nI (t )

(t) tan -1

(1.123)
(1.124)

Follow a similar procedure , we have


r
r 2 A2 2 Ar cos
f R , ( r, )
exp(
)
2
2
2
2
R and are dependent.

59

f R (r)

2
0

f R , ( r, )d

r
r 2 A2 2
Ar

exp(
) exp( 2 cos )d
2
2
0
2
2

(1.126)

The modified Bessel function of the first kind of zero


order is defined is (Appendix 3)
1 2
I0 ( x)
exp( x cos )d

2 0
Ar
r
r 2 A2
Ar
Let x 2 , f R(r) 2 exp(
) I0 ( 2 )
2

(1.127)
(1.128)

It is called Rician distribution.


60

r
A
Normalized , a

fV ( v ) f R ( r )
v2 a2
v exp(
) I 0 ( av )
2

(1.131)
(1.132)

Figure 1.23 Normalized Rician distribution

61

You might also like