LDPC码在基于类二分法SVM的MPPSK信号检测中的应用

发布时间：2018-04-28 10:16

本文选题：多元位置相移键控 + 类二分法SVM　；参考：《东南大学学报(自然科学版)》2014年03期

【摘要】：为了降低基于多分类支持向量机(SVM)的多元位置相移键控(MPPSK)系统的信号检测复杂度,同时提升误码率性能,提出一种新的类似于二分法的SVM多分类方法.然后在此基础上引入LDPC信道编码,考虑到SVM输出转化为多进制后验概率的复杂性,采用简单的位比特翻转法进行LDPC译码,并利用一种进制转化机制简化多进制LDPC译码.结果显示,类二分法SVM判决方法可以有效降低多分类SVM检测算法的复杂度,且其检测性能相较于已有的MPPSK信号检测算法显著提升;所采用的进制转化机制实现了基于SVM多分类判决的MPPSK系统中的多进制LDPC编解码;引入LDPC编码以后,MPPSK信号性能进一步显著提升.
[Abstract]:In order to reduce the signal detection complexity of multi-position phase-shift keying (MPPSK) system based on multi-classification support vector machine (SVM) and improve the BER performance, a new SVM multi-classification method similar to dichotomy is proposed. Based on this, LDPC channel coding is introduced. Considering the complexity of the conversion of SVM output to multiary posteriori probability, a simple bit flipping method is used to decode LDPC, and a new system is used to simplify the LDPC decoding. The results show that the SVM decision method can effectively reduce the complexity of the multi-class SVM detection algorithm, and its detection performance is significantly improved compared with the existing MPPSK signal detection algorithm. The binary conversion mechanism is used to realize the multiary LDPC coding and decoding in MPPSK system based on SVM multi-classification decision, and the performance of MPPSK signal is further improved after the introduction of LDPC coding.
【作者单位】：东南大学信息科学与工程学院;
【基金】：国家自然科学基金资助项目(61271204) 广东省重大科技专项资助项目(2010A080402003)
【分类号】：TN911.22

【参考文献】