The Optimization of Design for BICM System Based on Irregular LDPC Codes
|School||Harbin Institute of Technology|
|Course||Information and Communication Engineering|
|Keywords||irregular LDPC code bit-interleaved coded modulation system gaussian approximation density evolution constellation labeling|
The main issue of the future wireless communication is poor spectrum resources with the increasing demand of high speed and high quality communication. It is an efficient strategy to improve bandwidth efficiency, by combining irregular low-density parity-check(LDPC) codes of excellent error correction performance and bit-interleaved coded modulation(BICM) systems of high-bandwidth efficiency with high order modulation.At first, this dissertation summarizes foundational information of LDPC codes, including its definition, concept model, coding and decoding technologys, and the optimization of degree distributions for irregular LDPC codes: density evolution and GA (Gaussian Approximation) density evolution. With two methods of constructing Tanner graphs: Progressive edge-growth(PEG) method and random construction method compared, we know PEG method performances much better.We show that variable nodes of different degrees for irregular LDPC codes have different error correction performance.Then the dissertation introduces the concept model of BICM systems and the design paradigm of constellation labeling. We show that BICM systems based on joint iterative decoding gains 0.6 dB more than BICM systems without applying joint iterative decoding,though cost more iterative times and then some more decoding delay. For three typical constellation labeling methods: Gray labeling, MSP labeling , and SP labeling, Euclidean distance design criteria is given. For the AWGN channel, our simulations prove that Gray labeling performances best in the BICM systems, and that MSP labeling performances best in the BICM-ID systems.Finally, an improved mapping strategy for BICM systems with irregular LDPC codes was proposed. The mapping paradigm is to provide more coding protection through iterative decoding to bits that are less protected by modulation, with coding and modulation combined. By using Gaussian approximation (GA) density evolution algorithm, we mapped each equivalent subchannel with different degree distributions. Our simulations show that the proposed mapping strategy can improve the error-rate performance of conventional BICM mapping strategy based on irregular LDPC codes by more than one tenth of a decibel, without adding complexity.