7.2 Ergodic Random Processes
nal, such as the mean or the autocorrelation function, If two random processes are in-
volved, then we might also estimate the crosscorrelation function by time averaging as
follows.
= Rolo) = yim f xOYE +a va
Itis generally not possible to test whether or not a random process is ergodic in the
‘mean or the autocorrelation, since in practice the statistics of a random provess may not
be known, Consequently, it is often assumed that a process is ergodie in the mean or au-
‘ocorrelation or both. Such an assumption provides a method for obtaining an estimate
for the random process from one member function of the ensemble. The assumption that
4 random process is ergodic (in the mean or autocorrelation or both) is sometimes re-
ferred to as the ergodic hypothesis.
‘The properties ofthe autocorrelation and crosscorrelation functions estimated by time
averaging are the same as those for the same functions calculated by ensemble averaging.
Consider the sine wave random process with phase as the random variable, which is uniformly