Decoder-in-the-Loop: Genetic Optimization-Based LDPC Code Design

Abstract

LDPC code design tools typically rely on asymptotic code behavior and are affected by an unavoidable performance degradation due to model imperfections in the short length regime. We propose an LDPC code design scheme based on an evolutionary algorithm, the Genetic Algorithm (GenAlg), implementing a "decoder-in-the-loop" concept. It inherently takes into consideration the channel, code length and the number of iterations while optimizing the error-rate of the actual decoder hardware architecture. We construct short length LDPC codes (i.e., the parity-check matrix) with error-rate performance comparable to, or even outperforming that of well-designed standardized short length LDPC codes over both AWGN and Rayleigh fading channels. Our proposed algorithm can be used to design LDPC codes with special graph structures (e.g., accumulatorbased codes) to facilitate the encoding step, or to satisfy any other practical requirement. Moreover, GenAlg can be used to design LDPC codes with the aim of reducing decoding latency and complexity, leading to coding gains of up to 0.325 dB and 0.8 dB at BLER of 10 −5 for both AWGN and Rayleigh fading channels, respectively, when compared to state-of-the-art short LDPC codes. Also, we analyze what can be learned from the resulting codes and, as such, the GenAlg particularly highlights design paradigms of short length LDPC codes (e.g., codes with degree-1 variable nodes obtain very good results).

Performance analysis of the bioinspired method for optimizing irregular codes with a low density of parity checks

“…On the other hand, the high computational complexity of the optimization method based on the concept of "decoder in a loop" [14] limits the scope of its application to very short codes. Based on the study results, the proposed optimization 3 -optimized irregular code Fig.…”

“…In addition, the use of special mathematical tools in [12,13] with limitations for optimizing short codes leads to a deterioration in the characteristics of the obtained codes. An option of overcoming these difficulties is to apply the concept of "decoder in a loop" using a genetic algorithm when optimizing short codes [14]. A given approach takes into consideration the characteristics of the communication channel and meets the practical requirements for decoding.…”

NOMA for Next-Generation Massive IoT: Performance Potential and Technology Directions

“…Recently advances in artificial intelligence (AI) and machine learning have been found to be quite useful for designing good codes. In [13], it was proposed to use a genetic algorithm for constructing the parity check matrix of LDPC codes having short blocklengths, which exactly fits the small packet traffic pattern of massive IoT. In such a design approach, edges in the LDPC's Tanner graph are added or pruned via genetic trials, which departs from the classical design commencing from the optimization of the degree distribution of the parity-check matrices.…”

Deep Learning-Based Polar Code Design

“…In this work, we propose a deep learning-based polar code 1 construction algorithm. Following the same spirit as [19], [20], we learn the polar code (i.e., frozen and non-frozen indices) for a given decoder and a specific channel. This showed to be a rather simpler code construction method than explicitly finding the analytical solution for general channels while taking the decoder properties into account.…”