Improved sliding window decoding algorithm based on information reserved for spatially coupled LDPC codes

Abstract

Sliding window decoding (SWD) algorithm is an applicable and efficient way to decode spatially coupled low-density parity-check (SC-LDPC) codes. In the conventional SWD algorithm, the edges involved in the overlapping region between two adjacent decoding windows will maintain the updated log-likelihood-ratios (LLRs) in the current decoding window to the next decoding window. Error accumulation may occur under worse channel condition. To address up this problem, we propose an improved SWD algorithm by reserving more reliable LLRs to the next decoding window, denoted as IRSWD algorithm. Specifically, we first reserve the total LLRs for the variable nodes in the overlapping region of the current decoding window and compare them with the channel LLRs. Then choose more reliable LLRs as the initial LLR information and prior information for these variable nodes in the next decoding window. Simulation results show that the proposed IRSWD algorithm performs better than the conventional SWD algorithm for different SC-LDPC codes. Moreover, we also calculate the LLRs of the variable nodes in one decoding window under IRSWD algorithm and SWD algorithm respectively and the comparison results verify that the initial LLRs of the variable nodes in the overlapping region under the proposed IRSWD algorithm are more reliable than those under the conventional SWD algorithm.