Abstract
Abstract Low-density parity-check (LDPC) block codes have been shown to have near-capacity perfor- mance with iterative message-passing decoding and su±ciently long block length. However, most methods for designing LDPC block codes are based on random constructions; the lack of structures leads to serious disadvantages of high complexity in encoding and decoding. Therefore, in recent researches, codes with algebraic structures have been developed, among which quasi-cyclic LDPC (QC-LDPC) codes are an important class. LDPC convolutional codes (LDPC-CCs) are convolutional codes in nature but with sparse parity-check matri- ces. They can be encoded and decoded for arbitrary lengths of data with low complexity, suitable for applications such as packet-switching networks. In this thesis, we develop new constructions of QC-LDPC codes and LDPC-CCs with enlarged minimum distance. Simula- tion results show that codes by our construction can have better performance than previous codes at high signal-to-noise ratios.