PDD-Based Decoder for LDPC Codes With Model-Driven Neural Networks

作     者：Liu, Yihao Zhao, Ming-Min Wang, Chan Lei, Ming Zhao, Min-Jian 

作者机构：Zhejiang Univ Coll Informat Sci & Elect Engn Hangzhou 310027 Peoples R China IPCAN Zhejiang Prov Key Lab Informat Process ing Commun Hangzhou 310027 Peoples R China 

出 版 物：《IEEE COMMUNICATIONS LETTERS》 (IEEE Commun Lett)

年 卷 期：2022年第26卷第11期

页      面：2532-2536页

核心收录：

学科分类：0810[工学-信息与通信工程] 0808[工学-电气工程] 08[工学] 

基　　金：National Natural Science Foundation of China Zhejiang Provincial Natural Science Foundation of China [LQ20F010010] Defense Industrial Technology Development Program [JCKY2020210B021] Fundamental Research Funds for the Central Universities [226-2022-00195] 

主　　题：Decoding Iterative decoding Signal processing algorithms Neural networks Computational complexity Convex functions Convergence LDPC codes penalty dual decomposition deep learning deep unfolding model-driven 

摘      要：In this work, we develop a double-loop iterative decoding algorithm for low density parity check (LDPC) codes based on the penalty dual decomposition (PDD) framework. We utilize the linear programming (LP) relaxation and the penalty method to handle the discrete constraints and the over-relaxation method is employed to improve convergence. Then, we unfold the proposed PDD decoding algorithm into a model-driven neural network, namely the learnable PDD decoding network (LPDN). We turn the tunable coefficients and parameters in the proposed PDD decoder into layer-dependent trainable parameters which can be optimized by gradient descent-based methods during network training. Simulation results demonstrate that the proposed LPDN with well-trained parameters is able to provide superior error-correction performance with much lower computational complexity as compared to the PDD decoder.