Prior-Assisted Hierarchical ADMM Decoding for Punctured Globally Coupled LDPC Codes

Area of Science:

• Wireless communication systems engineering
• Information theory and coding
• Applied mathematics and optimization

Background:

• Future wireless networks demand channel coding with high reliability, low latency, and adaptability.
• Globally coupled low-density parity-check (GC-LDPC) codes offer enhanced error correction but face decoding complexity challenges.
• Conventional decoders struggle with the global constraints of GC-LDPC codes in heterogeneous scenarios.

Purpose of the Study:

• To propose a novel decoding framework for GC-LDPC codes that addresses complexity and reliability concerns.
• To develop a hierarchical decoding strategy that leverages local processing before global verification.
• To improve the efficiency and effectiveness of GC-LDPC decoding for advanced wireless systems.

Main Methods:

• A prior-assisted hierarchical alternating direction method of multipliers (ADMMs) decoding framework is introduced.
• The GC-LDPC parity-check structure is partitioned for parallel local ADMM decoding.
• A global fallback mechanism with fused reliability information and prior manipulation guides full-matrix ADMM decoding when necessary.

Main Results:

• The proposed local-to-global strategy reduces unnecessary global decoding iterations.
• The framework effectively preserves global consistency when local decoding is insufficient.
• Evaluations using bit error rate, frame error rate, and iteration counts demonstrate improved reliability and efficiency.

Conclusions:

• The developed framework offers an average-complexity-aware and reliability-aware decoding solution for GC-LDPC codes.
• This approach is suitable for advanced channel coding requirements in future wireless networks.
• The hierarchical ADMMs decoding method provides a practical advancement for high-performance wireless communications.