Design and Decoding LDPC Codes With Low Complexity
Date
Author
Institution
Degree Level
Degree
Department
Supervisor / Co-Supervisor and Their Department(s)
Examining Committee Member(s) and Their Department(s)
Citation for Previous Publication
Link to Related Item
Abstract
This thesis presents low complexity design and decoding schemes for low density parity check (LDPC) codes. First, we consider the iterative decoding of LDPC codes on multiple-input-multiple-output bit-interleaved coded modulation (MIMO-BICM) channels and two-way relay channels. More specifically, we study the log-likelihood ratio (LLR) calculation under MIMO-BICM channels when perfect channel information is known and LLR calculation for two-way relay channels when no channel information is known at the receiver. We propose the optimum piece-wise linear approximation in the sense of maximizing the achievable rate of the channel. Second, we introduce a novel “universal” LDPC code design method. We design universal LDPC codes based on our method and show that, compared to existing methods, a lager percentage of capacity is obtained. Then, we propose two conjectures about the extreme distributions under min-sum decoding based on numerical observations.