Connecting Spatially Coupled LDPC Code Chains

Codes constructed from connected spatially coupled low-density parity-check code (SC-LDPCC) chains are proposed and analyzed. It is demonstrated that connecting coupled chains results in improved iterative decoding performance. The constructed protograph ensembles have better iterative decoding thresholds compared to an individual SC-LDPCC chain and require less computational complexity per bit when operating in the near-threshold region. In addition, it is shown that the proposed constructions are asymptotically good in terms of minimum distance.

💡 Research Summary

This paper introduces a novel construction for spatially coupled low‑density parity‑check (SC‑LDPC) codes that connects two parallel (3,6) regular SC‑LDPC chains with two vertical “bridge” chains. The authors describe the protograph representation of the construction, where each bridge of length L/2 is attached to the horizontal chains at positions L/4 from the ends. The attachment is performed by linking low‑degree check nodes at the top of each bridge to variable nodes of the horizontal chains, thereby increasing the degree of those variable nodes from 3 to 4 while keeping all check‑node degrees ≤ 6.

Density‑evolution analysis is carried out for both the binary erasure channel (BEC) and the additive white Gaussian noise (AWGN) channel. For the BEC, the authors derive the standard erasure‑message update equations and show analytically that increasing variable‑node degree improves the erasure probability while increasing check‑node degree degrades it. Numerical results demonstrate that the connected ensemble S(3,6,L) achieves higher belief‑propagation (BP) thresholds than a single chain C(3,6,L′) of comparable rate. For example, with L = 24 (rate ≈ 0.444) the S‑ensemble’s BEC threshold exceeds that of a single chain of length 18 by roughly 0.03, moving the performance closer to channel capacity. Similar improvements are observed for the AWGN channel using discretized density evolution.

To address decoding complexity, the paper adopts the selective update schedule from