Layered PEXIT-based Optimization of Puncturing Patterns for 5G-LDPC Codes in Few-Iteration Decoding

In this paper, an innovative method to optimize the puncturing patterns of the 5G low-density parity-check codes, which leverages layered protograph-based extrinsic information transfer (PEXIT) analysis to evaluate decoding performance and the progressive multi-path search (PMPS) algorithm to optimizing puncturing patterns. Compared with existing methods, layered PEXIT analysis is faster and more accurate, while the PMPS algorithm effectively addresses the local optimum trap commonly encountered by greedy algorithms. Simulation results demonstrate that using optimized puncturing patterns significantly improve the few-iteration decoding performance, thereby enhancing the reliability of resource-constrained communication systems without introducing any additional complexity.