Two-Phase Globally Coupled Low-Density Parity Check Decoding Aided with Early Termination and Forced Convergence.

Area of Science:

• Information Theory
• Coding Theory
• Digital Communications

Background:

• Globally Coupled (GC) Low-Density Parity-Check (LDPC) codes are crucial for efficient data transmission.
• Traditional decoding algorithms can be computationally intensive, limiting efficiency.
• Enhancing decoding speed without compromising error correction is a key challenge.

Purpose of the Study:

• To improve the decoding efficiency of GC LDPC codes.
• To introduce Early Termination (ET) and Forced Convergence (FC) into a two-phase decoding algorithm.
• To analyze the performance and complexity trade-offs of the proposed methods.

Main Methods:

• Implemented a two-phase decoding algorithm integrating ET and FC techniques.
• Developed two variants of ET for local decoding: ET-1 (predictive termination) and ET-2 (iteration redistribution).
• Evaluated the combined ET-FC approach for GC LDPC codes.

Main Results:

• ET-1 significantly reduces decoding time complexity (up to 42%) with no performance loss.
• ET-2 maintains decoding time complexity while improving Bit Error Rate (BER) and Frame Error Rate (FER).
• The ET-2-FC combination offers a 25% complexity saving with improved BER (0.18 dB) and FER (0.23 dB).

Conclusions:

• Early Termination (ET) and Forced Convergence (FC) are effective for accelerating GC LDPC code decoding.
• ET-1 provides substantial complexity gains, while ET-2 offers performance improvements.
• The proposed ET-FC strategies enhance the practical applicability of GC LDPC codes in communication systems.