If I have bit sequence X at Alice and correlated sequence Y at Bob. How do I use LDPC to correct errors in Y such that X = Y. All examples I've seen about classical LDPC assume X is modulated, passed through a channel, and then demodulated at Bob to calculated LLR for the LDPC decorder. In my case I don't want to send X. Sending the syndrome of X is OK, but how do I go about using that information for belief propagation at Bob and how do I get LLR?
-
$\begingroup$ Who obtained Y and X and how? Was one derived from the other? $\endgroup$ – kodlu Feb 13 '17 at 2:30
-
$\begingroup$ @kdlu Yes, you can consider them as derived from each other since they are correlated. If it helps answer my question you can consider Y as slightly noisy version of X, for instance Y as output of BSC with X as input. $\endgroup$ – user3669561 Feb 13 '17 at 11:55