Syndrome-Based HARQ Schemes for Anytime Codes

Anytime code is a new class of error-correcting code with arbitrarily long coding memory (by design) so that its decoding performance can be improved by an expanding decoding window size (i.e., decoding delay). The anytime code has been shown to help stabilize an unstable system over noisy channels. Among practical anytime codes, spatially coupled repeat-accumulate (SC-RA) anytime codes have been proven to provide outstanding asymptotic performance while offering a simple encoding process. However, in the finite block-length regime, high erasure/error floors are observed for the SC-RA anytime codes even as the decoding window size (decoding delay) is enlarged. In this paper, to alleviate the high erasure/error floors problem, novel syndrome-based hybrid automatic repeat request (HARQ) schemes are designed for the SC-RA anytime codes over the binary erasure channel (BEC) and the binary input additive white Gaussian noise (BI-AWGN) channel, respectively. Simulation results show that the proposed syndrome-based HARQ schemes offer better decoding performance and spectral efficiency than the prior-art anytime code HARQ schemes.