Journal of Chongqing University of Posts and Telecommunications

EXIT charts for nonbinary LDPC codes

Yong Li

Published Jan 1, 2008• 0 citations

Abstract

Non-binary low-density parity-check(LDPC) codes defined over the finite field GF(q=2p) are an extension of binary LDPC codes and have demonstrated improved performance over binary LDPC codes with the same frame length and rate.This paper introduces the principle of extrinsic information transfer(EXIT) charts,analyzes the convergence behavior of non-binary LDPC codes on AWGN channel,and confirms the thresholds of non-binary LDPC codes,which provides the foundation for analysis and design of real communication system using EXIT tool in future work.Simulation result shows that the performance of quaternary LDPC codes is better than that of binary LDPC codes in the same channel,and validates the superiority of using EXIT chart for iterative research.

Related papers

EXIT charts for non-binary LDPC codes

G.J. Byers, Fambirai Takawira

In this paper extrinsic information transfer (EXIT) charts are proposed to design non-binary low-density parity-check (LDPC) codes for the AWGN channel. A new metric is presented to describe the mutual information of the non-binary messages. The a priori information is modelled using a Gaussian mixture distribution. Analytical expressions are given for the EXIT curves of the variable and check node decoders for both regular and irregular LDPC codes. The analytical expressions are shown to agree well with simulation results. Finally, by matching the variable and check node EXIT curves, it is shown that good nonbinary LDPC can be designed for the AWGN channel.

International Conference on CommunicationsAug 16, 200519 citations

EXIT chart analysis of regular and irregular LDPC convolutional codes on AWGN channel

Oulfa Laouar, Imed Amamra, Nadir Derouiche

Low-density parity-check (LDPC) codes are widely recognized for their excellent forward error correction, near-Shannon-limit performance, and support for high data rates with effective hardware parallelization. Their convolutional counterpart, LDPC convolutional codes (LDPC-CCs), offer additional advantages such as variable codeword lengths, unlimited parity-check matrices, and simpler encoding and decoding. These features make LDPC-CCs particularly suitable for practical implementations with varying channel conditions and data frame sizes. This paper investigates the performance of LDPC-CCs using the extrinsic information transfer (EXIT) chart, a graphical tool for analyzing iterative decoding. EXIT charts visualize mutual information exchange and help predict convergence behavior, estimate performance thresholds, and optimize code design. Starting with the EXIT chart principles for LDPC codes, we derived the mutual information functions for variable and check nodes in regular and irregular LDPC-CC tanner graphs. This involved adapting existing EXIT functions to the periodic parity-check matrix of LDPC-CCs. We compare regular and irregular LDPC-CC constructions, examining the impact of degree distributions and the number of periods in the parity-check matrix on convergence behavior. Our simulations show that irregular LDPC-CCs consistently outperform regular ones, and the EXIT chart analysis confirms that LDPC-CCs demonstrate superior bit error rate (BER) performance compared to equivalent LDPC block codes.

Bulletin of Electrical Engineering and InformaticsNov 22, 20241 citation

On the EXIT chart analysis of low-density parity-check codes

S.R. Kollu, Hamid Jafarkhani

The extrinsic information transfer (EXIT) charts are used to analyze the sum-product decoding algorithm of low-density parity-check (LDPC) codes. The codes are optimized by performing curve fitting on EXIT charts. This technique is developed for binary erasure channels based on area property and is extended to other channels. We show that on channels where the area property is not valid, the curve fitting approach is a sub-optimal method of optimizing for the best convergence threshold. On Gaussian channels, the messages from variable nodes to check nodes are approximated by a symmetric Gaussian distribution. We prove that the check node detector mutual information transfer function derived using this approximation is an upper bound. So the EXIT chart analysis predicts an optimistic value for the convergence threshold. We give an empirical lower bound for the check node detector mutual information transfer function. For few irregular LDPC codes designed for AWGN channel using density evolution, we have examined the accuracy of convergence thresholds predicted by EXIT charts using these bounds for the check node transfer function. The analysis of EXIT chart technique presented in this paper can be extended to other channels for which the distribution of channel messages can be approximated by a symmetric Gaussian/Gaussian mixture distribution.

GlobecomJan 1, 200514 citations

A comparison between BP, log-likelihood and max log-likelihood decoding algorithms of LDPC codes based on EXIT chart and EXIT trajectories methods

A. Refaey-Ahmed, Jean‐Yves Chouinard, Sébastien Roy +1 more

Low density parity check (LDPC) codes are known to achieve a performance very close to the Shannon capacity limit on Additive White Gaussian Noise (AWGN) and erasure channels. The extrinsic information transfer (EXIT) chart is a powerful method for analyzing iteratively decoded codes such as Turbo codes. EXIT charts were introduced as a method for representing how the mutual information between the decoder output and the transmitted bits changes over turbo decoding iterations. It is possible to apply the EXIT chart method to LDPC codes by treating the LDPC decoder as a concatenation of variable and check nodes. However the achieved convergence threshold values obtained with the EXIT charts are (plusmn0.1) dB away from the more precise results derived from the density evolution (DE) method. Subsequently, the EXIT trajectories method was proposed as an improved performance-analysis method for LDPC codes under belief propagation (BP) decoding and achieved more accurate convergence threshold values than the EXIT chart method. In this paper, the EXIT chart and the EXIT trajectories methods are proposed as analysis tools to compare between the BP, log likelihood and max log likelihood decoding algorithms for LDPC codes in terms of convergence thresholds. Furthermore, a comparison between EXIT chart, EXIT trajectories and DE methods is introduced for the mentioned LDPC codes. Simulations and numerical calculations on convergence thresholds for various ensembles of (dv, dc) regular LDPC codes for binary input AWGN channels are performed which results that the BP is the best decoding algorithm for LDPC code followed by the log likelihood and the max log likelihood algorithms, respectively. Also, the EXIT trajectory method yields more accurate results than EXIT charts which come very close to the DE method, while being more convenient numerically as a code research tool than the latter which require many Fourier transform operations.

International Conference on Advanced Technologies for CommunicationsOct 1, 20081 citation

EXIT-chart properties of the highest-rate LDPC code with desired convergence behavior

Masoud Ardakani, Terence Chan, Frank R. Kschischang

We consider uniparametric LDPC decoding schemes, i.e., the class of decoding algorithms for which an extrinsic information transfer (EXIT) chart analysis of the decoder is exact. We treat the general case of code design for a desired convergence behavior and provide necessary conditions and sufficient conditions that the EXIT chart of the maximum rate low-density parity-check code must satisfy. Our results generalize some of the existing results for the binary erasure channel: our results apply to all uniparametric decoding schemes and they apply to any desired convergence behavior.

IEEE Communications LettersJan 1, 200513 citations

A class of non-binary LDPC codes

Deepak Gilra

In this thesis we study Low Density Parity Check (LDPC) and LDPC like codes over \nnon-binary fields. We extend the concepts used for non-binary LDPC codes to generalize \nProduct Accumulate (PA) codes to non-binary fields. We present simulation results that \nshow that PA codes over GF(4) performs considerably better than binary PA codes at \nsmaller block lengths and slightly better at large block lengths. We also propose a trellis \nbased decoding algorithm to decode PA codes and show that its complexity is \nconsiderably lower than the message-passing algorithm. \nIn the second part of the thesis we study the convergence properties of non-binary PA \ncodes and non-binary LDPC codes. We use EXIT-charts to study the convergence \nproperties of non-binary LDPC codes with different mean column weights and show why \ncertain irregularities are better. Although the convergence threshold predicted by EXIT-charts \non non-binary LDPC codes is quite optimistic we can still use EXIT-charts for \ncomparison between non-binary LDPC codes with different mean column weights.

Sep 30, 20041 citation