Analysis and Design of Ultra-Reliable Short Blocklength Analog Fountain Codes
Ke Zhang,Jian Jiao,Zixuan Huang,Bowen Feng,Shaohua Wu,Bin Cao,Qinyu Zhang
DOI: https://doi.org/10.1109/vtcfall.2018.8690673
2018-01-01
Abstract:Machine-to-Machine (M2M) communications are expected to support extremely harsh requirements on both latency and reliability, which is characterized by the ultra-reliable, low-latency coding (uRLLC) technology in physical layer. In this paper, motivated by the recent development on the finite-blocklength information theory, we propose an ultra-reliable short blocklength analog fountain code (AFC) for M2M communications. First, we use the extrinsic information transfer (EXIT) chart to analyze the AFC compressive sensing belief propagation (CS-BP) decoding algorithm, by tracking the mutual information of AFC CS-BP decoding process, which related to the channel dispersion for the short blocklength AFC. Then, based on the EXIT chart analysis, we propose a Weight-set optimization progressive edge-growth (WO-PEG) encoding algorithm for the short blocklength AFC. Simulation results show that the proposed WO-PEG AFC scheme can effectively improve block error rate (BLER) in the short blocklength regime.