ROHINI COLLEGE OF ENGINEERING AND TECHNOLOGY
4.1 AUTO CORRELATION FUNCTION
DEFINITION
Let {X(t)} be a random process. Then the auto correlation function of {Y(t)} is
defined by RXX ( t1,t2 ) = E [ X(t1)X(t2)] (i.e) The auto correlation function of
{X(t)} is the expected value of product of two samples of {X(t)}.
Properties of Auto Correlation Function
Property 1 : Auto Correlation function is even (or) Prove that 𝑹𝑿𝑿 (−𝝉) =
𝑹𝑿𝑿 (𝝉)
Proof:
𝑅𝑋𝑋 (𝜏) = 𝐸 [𝑋 (𝑡 + 𝜏)𝑋(𝑡)]
𝑅𝑋𝑋 (−𝜏) = 𝐸 [𝑋 (𝑡 − 𝜏)𝑋(𝑡)]
= 𝐸 [𝑋 (𝑡)𝑋(𝑡 − 𝜏)]
𝑅𝑋𝑋 (−𝜏) = 𝑅𝑋𝑋 (𝜏)
∴ 𝑅𝑋𝑋 (𝜏) is an even function
Property 2 : Prove that 𝑹𝑿𝑿 (𝟎) = E[ X2(t)]
Proof:
𝑅𝑋𝑋 (𝜏) = 𝐸 [𝑋 (𝑡 + 𝜏)𝑋(𝑡)]
By taking 𝜏 = 0, We get
𝑅𝑋𝑋 (0) = 𝐸 [𝑋 (𝑡)𝑋(𝑡)] = E[ X2(t)]
∴ 𝑅𝑋𝑋 (0) = E[ X2(t)]
MA8451- PROBABILITY AND RANDOM PROCESSES
ROHINI COLLEGE OF ENGINEERING AND TECHNOLOGY
Property 3 : Prove that |𝑹𝑿𝑿 (𝝉)| ≤ 𝑹𝑿𝑿 (𝟎) (or) Prove that the maximum
value of 𝑹𝑿𝑿 (𝝉) 𝒊𝒔 𝑹𝑿𝑿 (𝟎)
Proof :
𝑅𝑋𝑋 (𝜏) = 𝐸 [𝑋 (𝑡 + 𝜏)𝑋(𝑡)]
2
[𝑅𝑋𝑋 (𝜏)]2 = [𝐸 [𝑋(𝑡 + 𝜏)𝑋(𝑡)]]
≤ 𝐸 [𝑋2 (𝑡 + 𝜏)]𝐸[𝑋2 (𝑡)] by Schwartz inequality
= 𝑅𝑋𝑋 (0)𝑅𝑋𝑋 (0)
∴ |𝑅𝑋𝑋 (𝜏)|2 ≤ [𝑅𝑋𝑋 (0)]2
|𝑅𝑋𝑋 (𝜏)|2 ≤ [𝑅𝑋𝑋 (0)]2
|𝑅𝑋𝑋 (𝜏)| ≤ 𝑅𝑋𝑋 (0)
Property 4 : If X(t) is a stationary process and it has no periodic component
,then 𝝁𝑿 = √ 𝐥𝐢𝐦 𝑹𝑿𝑿 (𝝉)
𝝉→∞
Proof : Since {X(t)} is a stationary process, then as 𝜏 → ∞, 𝑋(𝑡)𝑎𝑛𝑑 𝑋(𝑡 + 𝜏) are
independent and 𝜇𝑋 = 𝐸 [𝑋 (𝑡)]
𝑅𝑋𝑋 (𝜏) = 𝐸 [𝑋 (𝑡 + 𝜏)𝑋(𝑡)]
lim 𝑅𝑋𝑋 (𝜏) = lim 𝐸 [𝑋(𝑡 + 𝜏)𝑋(𝑡)]
𝜏→∞ 𝜏→∞
= lim 𝐸 [𝑋(𝑡 + 𝜏)]𝐸[𝑋(𝑡)]
𝜏→∞
= lim 𝜇𝑋 𝜇𝑋 = 𝜇𝑋 2
𝜏→∞
⇒ 𝜇𝑋 = √ lim 𝑅𝑋𝑋 (𝜏)
𝜏→∞
∴ 𝐸 [𝑋 (𝑡)] = √ lim 𝑅𝑋𝑋 (𝜏)
𝜏→∞
MA8451- PROBABILITY AND RANDOM PROCESSES
ROHINI COLLEGE OF ENGINEERING AND TECHNOLOGY
PROBLEMS UNDER PROPERTIES OF AUTO CORRELATION
FUNCTION
1. Check whether the following are valid auto correlation functions:
(i) 𝐟(𝛕) = 𝐀𝐜𝐨𝐬𝛚𝛕 𝐰𝐡𝐞𝐫𝐞 𝐀 𝐚𝐧𝐝 𝛚 𝐚𝐫𝐞 𝐩𝐨𝐬𝐢𝐭𝐢𝐯𝐞 𝐜𝐨𝐧𝐬𝐭𝐚𝐧𝐭𝐬
(ii) 𝐠(𝛕) = 𝐀𝐬𝐢𝐧𝛚𝛕 𝐰𝐡𝐞𝐫𝐞 𝐀 𝐚𝐧𝐝 𝛚 𝐚𝐫𝐞 𝐩𝐨𝐬𝐢𝐭𝐢𝐯𝐞 𝐜𝐨𝐧𝐬𝐭𝐚𝐧𝐭𝐬
|𝝉|
(iii) 𝒉(𝝉) = 𝟏 − where T is a constant.
𝑻
Solution :
(i) f(τ) = Acosωτ
f(−τ) = Acos(−ωτ)
= Acosωτ = f(τ)
f(−τ) = f(τ)
∴ f(τ) is the even function.
Hence f(τ) is the ACF of a process.
(ii) g(τ) = Asinωτ
g(−τ) = Asin(−ωτ)
= −Asinωτ = −g(τ)
g(−τ) = −g(τ)
∴ g(τ) is an odd function.
MA8451- PROBABILITY AND RANDOM PROCESSES
ROHINI COLLEGE OF ENGINEERING AND TECHNOLOGY
Hence g(τ) is not a ACF of a process.
|𝜏|
(iii) ℎ(𝜏) = 1 −
𝑇
|−𝜏|
ℎ(−𝜏) = 1 −
𝑇
|𝜏|
=1− = ℎ (𝜏)
𝑇
ℎ(−𝜏) = ℎ(𝜏)
∴ h(τ) is the even function.
Hence h(τ) is the ACF of a process.
𝟏
2. Let {X(t)} be a WSS process with 𝑹(𝝉) = 𝟐𝟓 + . Find the mean, mean
𝟏+𝟔𝝉𝟐
square value and variance of {X(t)}
Solution:
The mean value of {X(T)} is given by
E[X(t)] = √ lim 𝑅𝑋𝑋 (𝜏)
𝜏→∞
1
= √ lim (25 + )
𝜏→∞ 1+6𝜏2
= √25 + 0 = √25
∴ 𝑀𝑒𝑎𝑛 𝐸 [𝑋 (𝑡)] = 5
The mean square value of {X(t)} is given by
1
E[ X2(t)] = 𝑅𝑋𝑋 (0) = 25 + = 25 + 1 = 26
1+6(0)
The Variance of {X(t)} is given by
Var[X(t)] = E[ X2(t)] – E[X(t)]2
MA8451- PROBABILITY AND RANDOM PROCESSES
ROHINI COLLEGE OF ENGINEERING AND TECHNOLOGY
= 26 – 52 = 26 – 25 = 1
Variance 𝜎 2 = 4
3. A Stationary process has an auto correlation function is given by 𝑹(𝝉) =
𝟐𝟓𝝉𝟐 +𝟑𝟔
. Find mean ,mean square value and variance of { X(t)}
𝟔.𝟐𝟓𝝉𝟐 +𝟒
Solution :
The mean value of {X(T)} is given by
E[X(t)] = √ lim 𝑅𝑋𝑋 (𝜏)
𝜏→∞
25𝜏2 +36
= √ lim ( )
𝜏→∞ 6.25𝜏2 +4
36
𝜏2 (25+ 2 ) 25
𝜏
√ lim 2 4 =√ = √4 = 2
𝜏→∞ 𝜏 (6.25+ 2 ) 6.25
𝜏
∴ 𝑀𝑒𝑎𝑛 𝐸 [𝑋 (𝑡)] = 2
The mean square value of {X(t)} is given by
E[ X2 (t)] = RXX(0)
36
= =9
4
The variance of {X(t)} is given by
Var{X(t)} = E[ X2(t)] – E[X(t)]2
= 9 – 22 = 9 - 4 = 5
Variance 𝜎 2 = 5
4. Let {X(t)} be a stationary process with 𝑹(𝝉) = 𝟐 + 𝟔𝒆−𝟑|𝝉| . Find the mean
and variance of {X(t)}
MA8451- PROBABILITY AND RANDOM PROCESSES
ROHINI COLLEGE OF ENGINEERING AND TECHNOLOGY
Solution :
The mean value of {X(T)} is given by
E[X(t)] = √ lim 𝑅𝑋𝑋 (𝜏)
𝜏→∞
= √ lim (2 + 6𝑒 −3|𝜏| )
𝜏→∞
√2 + 6𝑒 −|∞| = √2 + 0 = √2
∴ 𝑀𝑒𝑎𝑛 𝐸 [𝑋 (𝑡)] = √2
The mean square value of {X(t)} is given by
E[ X2 (t)] = RXX(0)
= 2 + 6𝑒 −|0| = 2 + 6 = 8
The variance of {X(t)} is given by
Var{X(t)} = E[ X2(t)] – E[X(t)]2
= 8 – √2 2 = 8 - 2 = 6
Variance 𝜎 2 = 6
5. If {X(t)} is a WSS process with auto correlation function 𝑹𝑿𝑿 (𝝉) =
𝟗
𝟒𝒆−𝟐|𝝉| + . Find the mean and variance of Y = X(4) – X(2)
𝟒𝝉𝟐 +𝟑
Solution :
9
Given 𝑅 (𝜏) = 4𝑒 −2|𝜏| + …….(1)
4𝜏2 +3
First we have to find the mean and variance of {X(t)}
The mean value of {X(t)} is given by
E[X(t)] = √ lim 𝑅𝑋𝑋 (𝜏)
𝜏→∞
MA8451- PROBABILITY AND RANDOM PROCESSES
ROHINI COLLEGE OF ENGINEERING AND TECHNOLOGY
9
= √ lim (4𝑒 −2|𝜏| + )
𝜏→∞ 4𝜏2 +3
= √0 + 0 = √0
∴ 𝑀𝑒𝑎𝑛 𝐸 [𝑋 (𝑡)] = 0 … … . (2)
The mean square value of {X(t)} is given by
E[ X2 (t)] = RXX(0)
= 4 + 3 = 7 from (1)
The variance of {X(t)} is given by
Var{X(t)} = E[ X2(t)] – E[X(t)]2
=7–02=7
Variance 𝜎 2 = 7……..(3)
We have to find mean and variance of Y = X(4) – X(2)
(i) Mean of Y = E[Y] = E[ X(4) – X(2)]
= E[X(4)] – E [ X(2)]
=0 - 0=0
(ii) Var(Y) = Var [X(4) – X(2)]
= Var[X(4)] + Var[x(2)] – 2Cov[ X(4) . X(2)]
= 7 + 7 – 2 Cov (4,2)…………….(4)
We know that Cov ( 4,2 ) = RXX (4,2) – E[X(4)] E[X(2)]
= RXX (4 - 2) – 0 = RXX (2) From (2)
9
= 4e-2(2) +
16+3
9
= 4e-2(2) + from (1)
19
MA8451- PROBABILITY AND RANDOM PROCESSES
ROHINI COLLEGE OF ENGINEERING AND TECHNOLOGY
= 0.546
(4) ⇒ Var (Y) = 14 – 2(0.546)
= 14 – 1.092
∴ Var(Y) = 12.908
6. Suppose that { X(t)} is a WSS process with ACF R( t1, t2 ) = 𝟗 +
𝟒 𝒆−𝟎.𝟐|𝒕𝟏 −𝒕𝟐| .Detemine the mean,variance and the covariance of the RV’S
Z = X(5) and W = X(8).
Solution :
Given 𝑅 (𝜏) = 9 + 4 𝑒 −0.2|𝜏|
We have to find the mean and variance {X(t)}
(i) The mean value of {X(t)} is given by
E[X(t)] = √ lim 𝑅(𝜏)
𝜏→∞
= √ lim [9 + 4 e − 0.2|𝜏|]
𝜏→∞
= √9 + 4𝑒 −∞
= √9 + 0
∴ 𝑀𝑒𝑎𝑛 𝐸 [𝑋 (𝑡)] = 3 … … . (1)
(ii) The mean square value of {X(t)} is given by
E [ X2(t)] = R(0)
= 9 + 4 e0 = 13.
The variance of {X(t)} is given by
2
Var[X(t)] = 𝜎𝑋(𝑡) = E[ X2(t)] – E[X(t)]2
= 13 – 32 = 4……….(2)
(iii) Given Z = X(5) ; W = X(8)
Mean of Z = E(Z) = E[X(5)] = 3 from(1)
MA8451- PROBABILITY AND RANDOM PROCESSES
ROHINI COLLEGE OF ENGINEERING AND TECHNOLOGY
Mean of W = E(W) = E[X(8)] = 3 from(1)
Var(Z) = Var[X(5)] = 4 from (2)
Var (W) = Var[X(8)] = 4 from (2)
Cov(Z,W) = E[ ZW] – E[Z]E[W]
= E[X(5) X(8)] – 3 X 3
= R( 5,8) - 9 = 9 + 4 e−0.2|−3| − 9
= 4 e-0.6
7. If [X(t)] is a wide sense stationary process with auto correlation function
R(𝝉) =A𝒆−𝒂|𝝉| , Determine the second order moment of the RV X(8) – X(5)
Sol:
Given R(𝜏)= A𝑒 −𝑎|𝜏| ______(1)
E[𝑋2 (𝑡)] = R(0) = A𝑒 0 = A ….. ….. ….. (2)
The second order moment of X(8) - X(5) is given by
=E[(𝑋(8) − 𝑋(5))2 ]
=E[𝑋2 (8)] + 𝑋 2 (5) - 2X(8)X(5)]
=E[𝑋2 (8)] + E[𝑋 2 (5)] - 2E[X(8)X(5)]
= A + A - 2𝑅𝑋𝑋 (8,5) from(2)
= 2A - 2𝑅𝑋𝑋 (8,5)
= 2A - 2𝑅𝑋𝑋 (3)
= 2A - 2𝐴𝑒 −𝛼|3| from(1)
=2A(1 - 𝑒 −3𝛼 )
MA8451- PROBABILITY AND RANDOM PROCESSES
ROHINI COLLEGE OF ENGINEERING AND TECHNOLOGY
8. If {X(t)} is a WSS process, then prove that 𝑬[𝑿(𝒕 + 𝝉) − 𝑿(𝒕)]𝟐 =
𝟐[𝑹𝑿𝑿 (𝟎) − 𝑹𝑿𝑿 (𝝉)]
Proof :
𝐸 [𝑋(𝑡 + 𝜏) − 𝑋(𝑡)]2 = 𝐸 [𝑋2 (𝑡 + 𝜏) + 𝑋2 (𝑡) − 2 𝑋( 𝑡 + 𝜏)𝑋(𝑡)]
= 𝐸 [𝑋2 (𝑡 + 𝜏)] − 𝐸[𝑋 2 (𝑡)] − 2𝐸[ 𝑋( 𝑡 + 𝜏)𝑋 (𝑡)]
= 𝑅𝑋𝑋 (0) + 𝑅𝑋𝑋 (0) − 2𝑅𝑋𝑋 (𝜏)
= 2𝑅𝑋𝑋 (0) − 2𝑅𝑋𝑋 (𝜏)
𝐸[𝑋(𝑡 + 𝜏) − 𝑋(𝑡)]2 = 2[𝑅𝑋𝑋 (0) − 𝑅𝑋𝑋 (𝜏)]
9. Suppose that {X(t}is a WSS process with mean that 𝝁𝑿 ≠0 and that Y(t) is
defined by Y(t)= X(t+𝝉) - X(t), where 𝝉 > 0 is a constant. Show that the
mean of [Y(t)] is zero for all value of t and the variance of [Y(t)] is given
by 𝝈𝟐𝒀 = 2[𝑹𝑿𝑿 (0) - 𝑹𝑿𝑿 (𝝉)]. Is [Y(t)] a WSS process?
Sol:
Given: Y(t) = X(t +𝜏) - X(t)
Given {X(t)} is a WSS process.
∴(1) E[X(t)] = constant = 𝜇𝑋 and
∴(2) 𝑅𝑋𝑋 (𝑡1,𝑡2 ) is a function of 𝜏.
(i) Mean of Y(t) = E[Y(t)]
=E[X(t +𝜏) - X(t)]
=E[X(t + 𝜏)] - E[X(t)]
=𝜇𝑋 - 𝜇𝑋 = 0
MA8451- PROBABILITY AND RANDOM PROCESSES
ROHINI COLLEGE OF ENGINEERING AND TECHNOLOGY
2
(ii) 𝜎𝑌(𝑡) = Var[Y(t)]
= Var[X(t +𝜏) -X(t)]
= Var[X(t + 𝜏)] + Var[X(t)] - 2 Cov[X(t + 𝜏) X(t)]
= 𝜎𝑋2 (t) +𝜎𝑋2 (t) - 2{E[(X(t +𝜏)X(t)] -E[X(t+𝜏 )]E[X(t)]}
= 2𝜎𝑋2 (t) - 2[𝑅𝑋𝑋 (𝜏)- 𝜇𝑋 𝜇𝑋 ]
= 2[𝜎𝑋2 (t) - 𝑅𝑋𝑋 (𝜏)+𝜇𝑋2
=2[E[𝑋2 (𝑡)] - [𝐸[𝑋(𝑡)]]2] - 𝑅𝑋𝑋 (𝜏) + 𝜇𝑋2 ]
=2[E(𝑋2 (t) - 𝜇𝑋2 - 𝑅𝑋𝑋 (𝜏)+𝜇𝑋2 )]
=2[𝑅𝑋𝑋 (0)- 𝑅𝑋𝑋 (𝜏)]
(iii) 𝑅𝑌𝑌 (𝑡1,𝑡2 ) = E[Y(𝑡1)Y(𝑡2 )]
=E[[X(𝑡1+ 𝜏) - X(𝑡1)] [X(𝑡1 + 𝜏) - X(𝑡1)]]
=E[X(𝑡1+ 𝜏)X(𝑡2 + 𝜏) - X(𝑡1+ 𝜏)X(𝑡2 ) - X(𝑡1)X(𝑡1 + 𝜏) + X(𝑡1 ) X(𝑡2 )]
=E[X(𝑡1+ 𝜏)X(𝑡2 + 𝜏)] - E[X(𝑡1+ 𝜏)X(𝑡2 )] - E[X(𝑡1 )X(𝑡1+ 𝜏) ] + E[X(𝑡1) X(𝑡2 )]
=𝑅𝑋𝑋 (𝑡1+ 𝜏- 𝑡2 + 𝜏) - 𝑅𝑋𝑋 (𝑡1+ 𝜏- 𝑡2 ) -𝑅𝑋𝑋 (𝑡1- 𝑡2 - 𝜏) + 𝑅𝑋𝑋 (𝑡1 - 𝑡2 )
=𝑅𝑋𝑋 (𝜏) - 𝑅𝑋𝑋 (𝜏+ 𝜏) - 𝑅𝑋𝑋 (0) +𝑅𝑋𝑋 (𝜏)
𝑅𝑌𝑌 (𝑡1,𝑡2 ) = 2𝑅𝑋𝑋 (𝜏) - 𝑅𝑋𝑋 (2𝜏) - 𝑅𝑋𝑋 (0)
Which is a function of 𝜏.
Also we have E[Y(t)] = 0
∴ [Y(t)] a WSS process
MA8451- PROBABILITY AND RANDOM PROCESSES
ROHINI COLLEGE OF ENGINEERING AND TECHNOLOGY
MEAN ERGODIC THEOREM :
Let {X(t)} be a random process with constant mean 𝜇𝑋 . Then
̅̅̅𝑇̅) = 0.
{X(t)} is mean ergodic if lim 𝑉𝑎𝑟(𝑋
𝑇→∞
1. The auto correlation function for a stationary process {X(t)} is given by
𝟐
𝑹𝑿𝑿 (𝝉) = 9 +2𝒆−|𝝉|. Find the Mean of the random variable Y = ∫𝟎 𝑿(t)dt
and the variance of {X(t)}.
Sol:
The mean of {X(t)} is given by
E[X(t)] = √ 𝑙𝑖𝑚 𝑅𝑋𝑋 (𝜏)
𝜏→∞
= √ 𝑙𝑖𝑚 [9 + 2𝑒 −|𝜏|] = √9 = 3
𝜏→∞
The mean Square value of {X(t)} is given by
E[𝑋2 (𝑡)] = 𝑅𝑋𝑋 (0) = 9+2 = 11
The variance if {X(t)} is given by
Var[X(t)] = 𝜎𝑋2 = E[𝑋2 (𝑡)] - [𝐸[𝑋(𝑡)]]2 = 11 - 9 = 2
MA8451- PROBABILITY AND RANDOM PROCESSES
ROHINI COLLEGE OF ENGINEERING AND TECHNOLOGY
Next, we find the mean of Y as
2
E[Y] = E[∫0 𝑋(t) dt]
2
=∫0 𝐸 [X(t)] dt
2
= ∫0 3 dt
= 3[𝑡]20
=6
2. The auto correction function of for a stationary process [X(t)] is given
𝟏
by 𝑹𝑿𝑿 (𝝉) = 1 + 𝒆−𝟐|𝝉|. Find the mean and variance of S = ∫𝟎 𝑿(t)dt.
Sol.
X(t) is defined in (0,1). ∴T=1
Given R(𝜏) = 1+𝑒 −2|𝜏|
The mean of [x(t)] is given by
E[X(t)] = √ 𝑙𝑖𝑚 𝑅(𝜏) = √ 𝑙𝑖𝑚 [1 + 𝑒 −2|𝜏| ] = 1
𝜏→∞ 𝜏→∞
1
Given S = ∫0 𝑋(t)dt.
We have to Find the mean and variance of S
(i) The mean of S is given by
1 1
E[S] = ∫0 𝐸 [X(t)]dt = ∫0 (1)dt = [𝑡]10 = 1-0
MA8451- PROBABILITY AND RANDOM PROCESSES
ROHINI COLLEGE OF ENGINEERING AND TECHNOLOGY
∴E[S] = 1
(ii) To compute variance of S.
1 𝑇 1
̅̅̅
𝑋𝑇̅= ∫0 𝑋(t) dt = ∫0 𝑋(t)dt = S [∵ T = 1]
𝑇
̅̅̅𝑇̅) = 1 ∫𝑇 (1-|𝜏|) 𝐶𝑋𝑋 (𝜏)𝑑𝜏
Var(𝑋 Here T = 1
−𝑇
𝑇 𝑇
1
=∫−1.(1-|𝜏|) 𝐶𝑋𝑋 (𝜏) 𝑑𝜏 ….. …… …… (1)
𝐶𝑋𝑋 (𝜏) = 𝑅𝑋𝑋 (𝜏) - E[X(𝑡1)] E[X(𝑡2 )]
= 1 + 𝑒 −2|𝜏|- 1 x 1 [∵E[X(t)] = 1]
= 𝑒 −2|𝜏|
̅̅̅𝑇̅) = ∫1 (1-|𝜏|)𝑒 −2|𝜏|𝑑𝜏= 2 ∫1.(1-|𝜏|)𝑒 −2|𝜏|𝑑𝜏
(1) ⇒ Var(𝑋 −1 0
1 𝑒 −2𝜏 𝑒 −2𝜏
= 2 ∫0 (1-𝜏)𝑒 −2𝜏 𝑑𝜏 = 2[(1-𝜏)[ ] - (-1) ( )]10
2 (−2)2
𝑒 −2 1 1 𝑒 −2 1 𝑒 −2 1
= 2[ 0 + + - ] = 2[ + ]= +
4 2 4 4 4 2 2
̅̅̅𝑇̅) = 1(1 + 𝑒 −2) = 0.5677
Var(𝑋
2
Since S = ̅̅̅
𝑋𝑇̅, Var(S) = 0.5677
MA8451- PROBABILITY AND RANDOM PROCESSES
ROHINI COLLEGE OF ENGINEERING AND TECHNOLOGY
𝟏𝟎
3. If S = ∫𝟎 𝑿(t) dt, Find the mean and variance of S if E[X(t)] = 8 and
𝑹𝑿𝑿 (𝝉) = 64 + 𝟏𝟎𝒆−𝟐|𝝉|
Sol:
X(t) is defined in (0,10) 𝑠𝑖𝑛𝑐𝑒 T = 10
10
Given: S = ∫0 𝑋(t) dt ; E[X(t)] = 8 ; 𝑅𝑋𝑋 (𝜏) = 64 + 10𝑒 −2|𝜏|
10 10
(i) Mean of S, E(S) = ∫0 𝐸 [𝑋(𝑡)] 𝑑𝑡 = ∫0 8 𝑑𝑡 = 8[𝑡]10
0 = 80
(ii) To compute Variance of ̅̅̅
𝑋𝑇̅
1 10 𝑆
̅̅̅
𝑋𝑇̅ ∫ X(t) dt =
10 0 10
̅̅̅𝑇̅ ) = 1 ∫𝑇 .(1-|𝜏|) 𝐶𝑋𝑋 (𝜏)𝑑𝜏….. ….. …… (1)
Var(𝑋 −𝑇
𝑇 𝑇
𝐶𝑋𝑋 (𝜏) = 𝑅𝑋𝑋 (𝜏) - E[X(𝑡1)] E[X(𝑡2 )]
=( 64 + 10𝑒 −2|𝜏|) - ( 8 x 8)
=10𝑒 −2|𝜏|
1 10 |𝜏| 10 |𝜏|
̅̅̅𝑇̅ ) =
(1) ⇒ Var(𝑋 ∫ .(1- 10) 10𝑒 −2|𝜏|𝑑𝜏 = ∫−10(1- 10) 𝑒 −2|𝜏|𝑑𝜏
10 −10
10 |𝜏| 10 𝜏
=2 ∫0 (1- ) 𝑒 −2|𝜏|𝑑𝜏 = 2 ∫0 (1- ) 𝑒 −2𝜏 𝑑𝜏
10 10
𝜏 𝑒 −2𝜏 −1 𝑒 −2𝜏 10
=2[(1 − )( ) −( ) ]
10 −2 10 (−2)2 0
1 1 1 𝑒 −20 +20 − 1
= 2[0 + 𝑒 −20 + - ] = 2[ ]
40 2 40 40
MA8451- PROBABILITY AND RANDOM PROCESSES
ROHINI COLLEGE OF ENGINEERING AND TECHNOLOGY
1
̅̅̅𝑇̅ )=
Var(𝑋 [𝑒 −20 + 19 ] = 0.95 …. …. ….. (1)
20
𝑆
We have ̅̅̅
𝑋𝑇̅ = ̅̅̅𝑇̅
⇒ 10𝑋
10
̅̅̅𝑇̅ )= 102 Var(𝑋
Var(S) = Var(10𝑋 ̅̅̅𝑇̅ )
= 100 x 0.95 From (1)
Var (S) = 95
4. The random binary transmission process {X(t)} is a wide sense process
|𝝉|
with zero mean and auto correlation function R(𝝉) = 1- Where T is a
𝑻
constant. Find the mean and variance of the time average of {x(t)} over
(0,T). Is {X(t)} mean-ergodic?
Sol: Given mean of {X(t)} is zero.
∴E[X(t)] = 0
The time average of {X(t)} is over (0, T) given by
1 𝑇
̅̅̅
𝑋𝑇̅ = ∫0 𝑋(t) dt
𝑇
The mean value of time average of {X(t)} is given by
̅̅̅𝑇̅ )= 1E [ 1 ∫𝑇[ 𝑋(𝑡)] 𝑑𝑡]
E(𝑋
𝑇 𝑇 0
1 𝑇
= ∫ 𝐸 [X(t)] dt
𝑇 0
∵E[X(t)] = 0
= 0
The variance of time average of [X(t)] is given by
1 𝑇 |𝜏|
̅̅̅𝑇̅ )=
Var(𝑋 ∫ (1 - ) 𝐶𝑋𝑋 𝑑𝜏…. …. …. (1)
𝑇 −𝑇 𝑇
MA8451- PROBABILITY AND RANDOM PROCESSES
ROHINI COLLEGE OF ENGINEERING AND TECHNOLOGY
|𝜏|
Given 𝑅𝑋𝑋 (𝜏) = 1- ;|𝜏| ≤ T
𝑇
𝐶𝑋𝑋 (𝜏) = 𝑅𝑋𝑋 (𝜏) - E[X(t)]E[X(t + 𝜏)]
=𝑅𝑋𝑋 (𝜏) - 0
|𝜏|
= 1- ;|𝜏| ≤ T
𝑇
̅̅̅𝑇̅ )= 1 ∫𝑇 ( 1 − |𝜏|)2𝑑𝜏
(1) ⇒ Var(𝑋
𝑇 −𝑇 𝑇
2 𝑇 |𝜏| 2
=
𝑇
∫0 ( 1 − 𝑇
) 𝑑𝜏
2 𝑇 𝜏
=
𝑇
∫0 ( 1 − 𝑇)2𝑑𝜏
3|𝜏|
2 (1− 𝑇 ) 𝑇
= [ 1 ]
𝑇 3(− ) 0
𝑇
2 2
= - [0-1] =
3 3
2
̅̅̅𝑇̅ ) = ≠ 0
𝑙𝑖𝑚 Var(𝑋
𝑇→∞ 3
∴ [𝑋(𝑡)] 𝑖𝑠 𝑛𝑜𝑡 𝑚𝑒𝑎𝑛 𝑒𝑟𝑔𝑜𝑑𝑖𝑐 𝑏𝑦 𝑚𝑒𝑎𝑛 𝑒𝑟𝑔𝑜𝑑𝑖𝑐 𝑡ℎ𝑒𝑜𝑟𝑒𝑚.
MA8451- PROBABILITY AND RANDOM PROCESSES
