An Early Termination Criterion for Stochastic LDPC Decoding
2016-06-27 18:57:15   来源：   评论：0 点击：

Abstract—Stochastic decoding is an excellent approach for Low-Density Parity-Check (LDPC) codes which are adopted in many communication standards, including 10GBASE-T, DVB-S2, WiMAX. In this paper, we first research some novel schemes of stochastic LDPC decoding, like Noise-Dependent Scaling (NDS), Edge Memories (EMs), Tracking forecast memories (TFMs) and majority-based tracking Forecast Memories (MTFMs). (7, 4) Hamming code and (2048, 1723) LDPC code from the IEEE 802.3an standard are used to simulate and analyze the advantages and disadvantages of these methods. Then, an early termination criterion is proposed as an efficient way to speed up the decoding procedure. Simulation results show that our method can decrease the computation complexity significantly with negligible performance loss.

 

Index Terms—Low-Density Parity-Check (LDPC), stochastic decoding, Edge Memories (EMs), low complexity, early termination criterion

Cite: Huiming Dai, Qiangze Chi, Hao Yan, Xin Zhang, and Dacheng Yang, "An Early Termination Criterion for Stochastic LDPC Decoding," Lecture Notes on Information Theory, Vol. 4, No. 1, pp. 7-12, June 2016. doi: 10.18178/lnit.4.1.7-12