Continuity and Differentiation of Random Processes

We know that the dynamic behavior of a system is described by differential

equations involving input to and output of the system. For example, the behavior
of an RC network is described by a first order linear differential equation with the
source voltage as the input and the capacitor voltage as the output. What happens
if the input voltage to the network is a random process?
ach reali!ation of the random process is a deterministic function and the
concepts of differentiation and integration can be applied to it. "ur aim is to
extend these concepts to the ensemble of reali!ations and apply calculus to
random process.
We discussed the convergence and the limit of a random sequence. #he continuity of the
random process can be defined with the help of convergence and limits of a random
process. We can define continuity with probability $, mean%square continuity, and
continuity in probability etc. We shall discuss the mean-square continuity and the
elementary concepts of corresponding mean-square calculus.
Mean-square continuity of a random process
Recall that a sequence of random variables { }
n
X
converges to a random variable X if
[ ]
&
lim '
n
n
E X X

and we write
l.i. m.
n
n
X X

( random process ( ) { }
X t
is said to be continuous at a point
'
t t
in the mean%square
sense if
( ) ( )
'
'
l.i. m.
t t
X t X t

or equivalently
( ) ( )
'
&
'
lim '
t t
E X t X t

1
]
Mean-square continuity and autocorrelation function
)$* ( random process ( ) { }
X t
is +, continuous at
'
t
if its auto correlation function
( )
$ &
,
X
R t t
is continuous at
' '
) , *. t t
Proof:
( ) ( ) ( ) ( ) ( ) ( ) ( )
( ) ( ) ( )
&
& &
' ' '
' ' '
&
, & , ,
X X X
E X t X t E X t X t X t X t
R t t R t t R t t
+ 1
]
+
-f ( )
$ &
,
X
R t t
is continuous at
' '
) , *, t t
then
( ) ( ) ( ) ( ) ( )
( ) ( ) ( )
' '
&
' ' ' '
' ' ' ' ' '
lim lim , & , ,
, & , ,
'
X X X
t t t t
X X X
E X t X t R t t R t t R t t
R t t R t t R t t

+ 1
]
+

)&* -f ( ) { }
X t
is +, continuous at
'
t
its mean is continuous at
'
. t
#his follows from the fact that
( ) ( ) ( ) ( ) ( ) ( )
& &
' '
E X t X t E X t X t 1 1
] ]
( ) ( ) ( ) ( )
' '
& &
' '
lim lim '
t t t t
E X t X t E X t X t

1 1 1
] ] ]
( ) EX t
is continuous at
'
. t
Example
Consider the random binary wave
. ) */ X t
discussed in xample . ( typical reali!ation
of the process is shown in Fig. below. #he reali!ation is a discontinuous function.
#he process has the autocorrelation function given by
$
) *
' otherwise

p
p X
T
T R

'

We observe that
) *
X
R
is continuous at '. #herefore,
) *
X
R
is continuous at all
.
p
T
$
$
'
) * X t
t
K
K
p
T
Example For a Wiener process ( ) { }
, X t
( ) ( )
( ) ( )
$ & $ &
, min ,
where is a constant.
, min ,
X
X
R t t t t
R t t t t t

#hus the autocorrelation function of a Wiener process is continuous everywhere implying
that a Wiener process is m.s. continuous everywhere. We can similarly show that the
0oisson process is m.s. continuous everywhere.
Mean-square differentiability
#he random process ( ) { }
X t
is said to have the mean%square derivative ( ) 1 X t
at a point
, t
provided
( ) ( ) X t t X t
t
+

approaches ( ) 1 X t
in the mean square sense as
' t . -n other words, the random process ( ) { }
X t
has a m%s derivative ( ) 1 X t
if
( ) ( )
( )
&
'
lim 1 '
t
X t t X t
E X t
t

+ 1

1

]
Remark
)$* -f all the sample functions of a random process ( ) X t
are differentiable, then the
above condition is satisfied and the m%s derivative exists.
Example
Consider the random%phase sinusoid { } ) * X t
given by
'
) * cos) * X t A w t +
where
'
and A w
are constants and
2 3', & 4.
#hen for each
,

) * X t
is differentiable. #herefore, the m.s. derivative is
' '
) * sin) * X t Aw w t +
M.S. Deriatie and !utocorrelation functions
#he m%s derivative of a random process ( ) X t
at a point t exists if
( )
&
$ &
$ &
,
X
R t t
t t

exists at the point
) , *. t t
(pplying the Cauchy criterion, the condition for existence of m%s derivative is
( ) ( ) ( ) ( )
$ &
&
$ &
, '
$ &
lim '
t t
X t t X t X t t X t
E
t t

+ + 1

1

]
xpanding the square and taking expectation results,
( ) ( ) ( ) ( )
( ) ( ) ( ) ( ) ( ) ( )
( ) ( ) ( ) ( )
$ &
$ &
&
$ &
, '
$ &
$ $ $ & & &
& &
, '
$ &
$ & $ &
$ &
lim
, , & , , , & ,
lim
, , , ,
&
t t
X X X X X X
t t
X X X X
X t t X t X t t X t
E
t t
R t t t t R t t R t t t R t t t t R t t R t t t
t t
R t t t t R t t t R t t t R t t
t t


+ + 1

1

]
+ + + + + + + + 1 1
+
1 1

] ]
+ + + + + 1

1

]
ach of the above terms within square bracket converges to
( )
$ &
&
$ &
$ &
,
,
X
t t t t
R t t
t t

1
1

]
if the
second partial derivative exists.
( ) ( ) ( ) ( ) ( ) ( ) ( )
$ &
$ &
&
& & &
$ & $ & $ & $ &
, '
$ & $ & $ & $ &
,
, , ,
lim &
'
X X X
t t
t t t t
X t t X t X t t X t R t t R t t R t t
E
t t t t t t t t


1 + + 1
+
1 1

] ]

#hus, ( ) { }
X t
is m%s differentiable at t if
( )
&
$ &
$ &
,
X
R t t
t t

exists at
) , * . t t
0articularly, if ( ) X t
is W,,,
( ) ( )
$ & $ &
,
X X
R t t R t t
,ubstituting
$ &
t t
, we get
( ) ( )
( ) ( )
( )
( )
& &
$ & $ &
$ & $ &
$ &
$ &
&
&
$
&
&
,
.
X X
X
X
X
R t t R t t
t t t t
dR t t
t d t
d R
d t
d R
d

#herefore, a W,, process ( ) X t
is m%s differentiable if ( )
X
R
has second derivative at
' .
Example
Consider a W,, process ( ) { }
X t
with autocorrelation function
( ) ( )
exp
X
R a
( )
X
R
does not have the first and second derivative at '. ( ) { }
X t
is not mean%square
differentiable.
Example #he random binary wave ( ) { }
X t
has the autocorrelation function
$
) *
' otherwise

p
p X
T
T R

'

( )
X
R
does not have the first and second derivative at '. #herefore, ( ) { }
X t
is not
mean%square differentiable.
Example For a Wiener process ( ) { }
, X t
( ) ( )
( )
( )
( )
$ & $ &
& &
&
&
&
&
$
&
&
$ &
$ &
$ &
, min ,
where is a constant.
if '
',
' other wise
if '
',
' if '
does not exist if if '
,
does not exist at ) ', '*
X
X
X
X
R t t t t
t t
R t
t
R t
t
t
t
R t t
t t
t t

<

'

<

>
'

#hus a Wiener process is m.s. differentiable nowhere.
Mean and !utocorrelation of t"e Deriatie process
We have,
( )
( ) ( )
( ) ( )
( ) ( )
( )
'
'
'
1 lim
lim
lim
1
t
t
X X
t
X
X t t X t
EX t E
t
EX t t EX t
t
t t t
t
t

For a W,, process ( ) ( ) 1 1 '

X
EX t t
as ( )
X
t
5 constant.
( ) ( ) ( )
( )
( ) ( )
( ) ( ) ( ) ( )
( ) ( )
( )
&
&
&
$ & 1 $ &
& & &
$
'
&
$ & & $ &
'
&
$ & & $ &
'
&
$ &
&
1 ,
lim
lim
, ,
lim
,
XX
t
t
X X
t
X
EX t X t R t t
X t t X t
EX t
t
E X t X t t X t X t
t
R t t t R t t
t
R
t t
t

+ 1
]

,imilarly we can show that

( ) ( )
( )
&
$ &
$ &
$ &
,
1 1
XX
R t t
EX t X t
t t

For a W,, process
( ) ( ) ( )
( )
$ & $ &
1
X
X
dR
EX t X t R t t
t d

and

( ) ( )
( )
( )
( ) ( )
( )
&
$ &
$ &
$ &
&
&
&
&
'
var
X
X
X
R t t
EX t X t
t t
d R
d
d R
X t
d

Mean Square #nte$ral
Recall that the definite integral )Riemannian integral* of a function ( ) ! t
over the interval
[ ]
'
, t t
is defined as the limiting sum given by
( ) ( )
'
$
'
'
lim
t
n
k k
n k
k
t
! d !

Where
' $ $
................
n n
t t t t t

< < < <

are partitions on the interval [ ]
'
, t t
and
$ k k k
t t
+

and [ ]
$
,
k k k
t t

.
For a random process ( ) . / X t
, the m%s integral can be similarly defined as the process
( ) . / " t
given by
( ) ( ) ( )
'
$
'
'
l.i .m.
t
n
k k
n k
k
t
" t X d X

Existence of M.S. #nte$ral
-t can be shown that the sufficient condition for the m%s integral ( )
'
t
t
X d

to
exist is that the double integral ( )
' '
$ & $ &
,
t t
X
t t
R d d

exists.
-f ( ) . / X t
is +.,. continuous, then the above condition is satisfied and the
process is +.,. integrable.
Mean and !utocorrelation of t"e #nte$ral of a %SS process
We have
( ) ( )
( )
'
'
'
'


) *
t
t
t
t
t
X
t
X
E" t E X d
EX d
d
t t

#herefore, if
',
X

( ) { }
" t
is necessarily non%stationary.
( )
( ) ( )
( ) ( )
( )
$ &
' '
$ &
' '
$ &
' '
$ & $ &
$ & $ &
$ & $ &
$ & $ &
, ) * ) *



"
t t
t t
t t
t t
t t
X
t t
R t t E" t " t
E X X d d
EX X d d
R d d

which is a function of
$ &
and . t t
Thus the integral of a WSS process is always non-stationary.
Remark #he nonstatinarity of the +.,. integral of a random process has physical
importance 6 the output of an integrator due to stationary noise rises unboundedly.
Example #he random binary wave ( ) { }
X t
has the autocorrelation function
Fig. )a* Reali!ation of a W,, process
) * X t

)b* corresponding integral
) * " t
$
) *
' otherwise

p
p X
T
T R

'

( )
X
R
is continuous at ' implying that ( ) { }
X t
is +.,. continuous. #herefore,
( ) { }
X t
is mean%square integrable.