学术活动
How Simple Can It be to Construct Good Codes
2015-04-25
供稿:科技处 点击次数:
主 讲 人: 马啸 教授(中山大学)
时 间: 2015年4月25日(周六)10:00-11:00
地 点: 必赢76net线路官网北二区教学楼 130 教室
主 办 单 位: 必赢76net线路官网数学科学学院
主讲人介绍:马啸,工学博士,中山大学软件学院院长;信息科学与技术学院教授、博士生导师(信息与通信工程);数学与计算科学学院兼职博导(应用数学)。主要研究信息论与信道编码技术及其在数字通信和数字存储系统中的应用,在《IEEE信息论汇刊》等国际重要刊物和IEEE信息论年会等重要国际会议上发表论文近百篇,其中一项成果被国外教授大篇幅写进专业教科书。2006年,获IEEE通信学会颁发的2005年度最佳论文奖(2005 Best Paper Award in Signal Processing and Coding for Data Storage by IEEE Comm. Soc.);2006年获首届微软青年教授奖;2007年入选教育部新世纪优秀人才计划;2008年成为广东省高校“千百十工程”省级培养对象;2012年主持国家重点基础研究发展计划(973计划)项目“高移动性宽带无线通信网络重点理论基础研究”课题之一。2014年至今,中国电子学会信息论分会副主任委员。
内 容 介 绍:In this talk, we present a simple approach to construct good codes with any given alphabet (moderate in size) and any code rate (rational one) to achieve a given target bit-error-rate (BER) over the additive white Gaussian noise (AWGN) channel. Firstly, we discuss about repetition (R) codes and single-parity-check (SPC) over groups. By time-sharing between the R code and the SPC code, we can construct codes with any given rate, which are referred to as the RSPC codes. The performance of this class of code can be obtained from the performances of the R code and the SPC code easily. Finally, by embedding the RSPC codes in the block Markov superposition transmission (BMST), we have a class of good codes-BMST-RSPC codes, which can approach any given target BER by choosing the encoding memory. A distinguished feature of the BMST-RSPC codes is the simple relation between the asymptotic coding gain and the encoding memory, which is helpful for designing the BMST-RSPC codes. Simulation results shows that the BMST-RSPC codes have a good performance (approaching the Shannon limit within 1 dB at the target BER) over a wide range of code rates.
时 间: 2015年4月25日(周六)10:00-11:00
地 点: 必赢76net线路官网北二区教学楼 130 教室
主 办 单 位: 必赢76net线路官网数学科学学院
主讲人介绍:马啸,工学博士,中山大学软件学院院长;信息科学与技术学院教授、博士生导师(信息与通信工程);数学与计算科学学院兼职博导(应用数学)。主要研究信息论与信道编码技术及其在数字通信和数字存储系统中的应用,在《IEEE信息论汇刊》等国际重要刊物和IEEE信息论年会等重要国际会议上发表论文近百篇,其中一项成果被国外教授大篇幅写进专业教科书。2006年,获IEEE通信学会颁发的2005年度最佳论文奖(2005 Best Paper Award in Signal Processing and Coding for Data Storage by IEEE Comm. Soc.);2006年获首届微软青年教授奖;2007年入选教育部新世纪优秀人才计划;2008年成为广东省高校“千百十工程”省级培养对象;2012年主持国家重点基础研究发展计划(973计划)项目“高移动性宽带无线通信网络重点理论基础研究”课题之一。2014年至今,中国电子学会信息论分会副主任委员。
内 容 介 绍:In this talk, we present a simple approach to construct good codes with any given alphabet (moderate in size) and any code rate (rational one) to achieve a given target bit-error-rate (BER) over the additive white Gaussian noise (AWGN) channel. Firstly, we discuss about repetition (R) codes and single-parity-check (SPC) over groups. By time-sharing between the R code and the SPC code, we can construct codes with any given rate, which are referred to as the RSPC codes. The performance of this class of code can be obtained from the performances of the R code and the SPC code easily. Finally, by embedding the RSPC codes in the block Markov superposition transmission (BMST), we have a class of good codes-BMST-RSPC codes, which can approach any given target BER by choosing the encoding memory. A distinguished feature of the BMST-RSPC codes is the simple relation between the asymptotic coding gain and the encoding memory, which is helpful for designing the BMST-RSPC codes. Simulation results shows that the BMST-RSPC codes have a good performance (approaching the Shannon limit within 1 dB at the target BER) over a wide range of code rates.