An Improved Data-Driven Decision Feedback Receiver via Deep Unfolding
Algorithm 2
The training algorithm of improved-DFR.
(1)
Input: Frame number , the received sampled signal set and the corresponding transmitted symbol sequence set , initial learning rate , frame size , training episode times , iteration times ,
(2)
Initialization: Randomly initialize , initialize Adam with learning rate ,
(3)
for to do
(4)
for to do
(5)
Randomly select in and the corresponding in . Transform to , transform to in format
