Braided Convolutional Self-orthogonal Codes with Double Sliding Window Decoding

In this paper, we investigate a class of braided convolutional codes (BCCs), where the component codes are convolutional self-orthogonal codes (CSOCs), called braided convolutional self-orthogonal codes. Compared to conventional BCCs, the advantages of braided CSOCs include the availability of several low-complexity decoding methods and the relative ease of extending these methods to high rates More specifically, to construct high-rate braided codes, it is necessary to use higher-rate component codes, which results in exponentially higher decoding complexity with conventional BCJR decoding, whereas the complexity of the CSOC decoding methods proposed here grows only linearly with rate. In particular, we introduce a double sliding window decoding method based on belief propagation (BP) for braided CSOCs, which exhibits good performance while maintaining low-complexity decoding for the higher rates required in many applications.