
DFE is simply the delayed copies of the transversal filter’s input signal.
Mathematically,
c
nþ1
= c
n
þ m e
n
X (18.10)
where
c
n
is the adapting coefficient vector in the nth iteration, e
n
is the corre-
sponding error signal and
X includes the delayed copies of the input signal which
are weighted by the coefficient vector
c
n
m is an adaptation parameter which is set
based on the desired adaptation speed, stability, and tolerable misadjustment for
different classes of optical channels. There are considerable variants from this
simple form to allow for one or more of further implementation simplicity,
stability, fast convergence, and/or tracking.
An LE using ...