by Ensemble Averaging
sense. This can be
th frequency wy. Such
Now imagine that the
ct sense stationary
tion coefficient fo
your result with tha
1 and x
od the autocorrelai
vionary in the wide
Se expectation with r=
x + 2) with resp
fori = 12
Sia(t) sin(t +
oth the mean fs
s nonstationary.
= for random proces
2, x(t) = 5. The
x(0) occurs with =
sses, namely, nos
fa set structure =
sses isthe subset
stationary random
random process
ON
ction for wide seams
correlation funciamy
cal descriptor a
64 Properties ofthe Autocorrelation Function 217
Figure 6.4 A set structure for tree classes of random processes
{Ne will assume thatthe autocorrelation function exists for every value of the argu-
ment, x, and is the same for all member functions of the ensemble
Property 6.4.1 We have already seen that the autocorrelation function at = O is the
erage power of the random process
Property 6.4.2 The autocorrelation function of a wide sense stationary real random
process is an even function; that is:
Ryx(t) = Rax( (6.4.1)
This latter property is established using the ensemble average definition for a wide
scase stationary real random process, as follows:
Rex(t) = EIX()X(t + 1)] = EIX(t— 7) X(]= Rex(—2) [64}
Scperty 6.4.3 The autocorrelation function for a wide sense stationary real random
process is maximum at the origin.
‘This property is established using the fact that for any two random variables, X and
me have
(E[XY])? = BEX*]ELY7] [6.4.3]
x(t) = {EIX(OX(t + 2)]
DP () IEDE(t +
RRx(0) 16.4.4)
IRux(t) | = Ruoc(0) [64.51
SSPey 6.44 IERix(0) = Rxx(T), then the autocorrelation function is periodic with
$2 T and the random process, X(), is said to be mean square periodic, thet is,
FXit + T) — X(0)"] = 0. The proof of this property is left as an exercise