基于多重差异化压缩感知的新型数字电视传输体制研究

Digital terrestrial multimedia/television broadcasting (DTMB) standard proposed by China, was officially approved by ITU in December 2011 as the fourth international digital television terrestrial broadcasting (DTTB) standard. The key technology of DTMB is time domain synchronous orthogonal frequency division multiplexing (TDS-OFDM), which achieves higher spectral efficiency and faster synchronization than classical OFDM techniques, but suffers from two problems: 1) The difficulty to support higher order modulations like 256QAM under deep frequency-selective fading channels; 2) The performance loss over fast fading channels. To solve those issues, we break through the conventional concept of "removing the interference if present", and propose a novel TDS-OFDM transmission architecture based on multiple perspective compressive sensing. Hereby, the distinct technical feature of TDS-OFDM, namely, the joint time-frequency domain processing, as well as the the multiple differentiated statistics of wireless channels, have been fully exploited. In this project, we try to solve three key scientific problems: 1) The optimal joint time-frequency domain processing under the constraint of high spectral efficiency; 2) The optimal training sequence design under the dual criteria of minimization of the mean squared error and minimization of the maximal cross-correlation; 3) The high-accuracy yet low-complexity compressive sensing signal reconstruction algorithms based on the multiple differentiated statistics of wireless channels. The achievements from this project will improve the spectral efficiency and the robustness of high-speed mobile reception for TDS-OFDM based broadcasting systems, and more importantly, let Chinese future DTTB evolution systems be capable of competing with other newly announed international DTTB standards.

2011年12月，中国数字电视地面多媒体广播标准（DTMB）被国际电信联盟正式采纳为国际标准。DTMB的基础调制理论时域同步正交频分复用（TDS-OFDM）具有高频谱效率和快速时域同步等显著优势,但现有体制存在长时延恶劣信道下难以支持256QAM等高阶调制、快速时变信道下性能恶化明显等问题。本项目将突破"有干扰就必须先消除"的传统思路，结合TDS-OFDM时频联合处理的鲜明技术特点，充分利用无线多径信道在统计特征上的多重差异性，提出基于多重差异化压缩感知的新型TDS-OFDM传输体制，通过解决TDS-OFDM在高频谱效率约束下的最优时频域联合处理、均方误差最小化和最大互相关最小化"双重准则"下训练序列的协同最优设计、基于无线信道多重差异化特征的高精度低复杂度信号重构等关键科学问题，为我国地面数字电视演进系统在频谱效率、高速移动接收等主要性能指标上赶超国际先进水平提供强有力的理论支撑。

针对中国数字电视地面多媒体广播标准（DTMB）的基础调制理论时域同步正交频分复用（TDS-OFDM）在长时延恶劣信道下难以支持256QAM等高阶调制、快速时变信道下性能恶化明显等问题，本项目突破“有干扰就必须先消除”的传统思路，结合TDS-OFDM时频联合处理的鲜明技术特点，充分利用无线多径信道在统计特征上的多重差异性，研究基于多重差异化压缩感知的新型TDS-OFDM传输体制，通过解决TDS-OFDM在高频谱效率约束下的最优时频域联合处理、均方误差最小化和最大互相关最小化“双重准则”下训练序列的协同最优设计、基于无线信道多重差异化特征的高精度低复杂度信号重构等关键科学问题，为我国地面数字电视演进系统在频谱效率、高速移动接收等主要性能指标上赶超国际先进水平提供强有力的理论支撑。项目执行期间，提出①基于压缩感知理论的多载波传输体制；②基于时频域协同优化的帧结构和信令设计；③基于广播信道特征和压缩感知理论的自适应信道估计；④基于无线信道时域稀疏特性的信道估计；⑤基于信道空时相关性的多输入多输出参数化信道估计；⑥FDD Massive MIMO系统中基于信道角度域稀疏特性的信道估计。在IEEE Transactions上发表SCI收录论文15篇；在其他SCI期刊上发表论文10篇；在ICC/Globecom等国际会议上发表EI收录论文10篇，申请国家发明专利6项；获IEEE ICC2013最佳论文奖、2016年IEEE Scott Helt Memorial Award（IEEE Transactions on Broadcasting年度唯一最佳论文奖）等奖项。本项目取得的基础理论成果推动了中国地面数字电视传输演进标准（DTMB-A）的研发。DTMB-A于2015年7月被正式列入国际电联ITU-R BT.1306建议书“数字地面电视广播的纠错、数据成帧、调制和发射方法”，成为数字电视国际标准，其频谱效率、高速移动接收等主要性能指标达到国际先进水平。总体而言，本项目超额完成了任务书规定的研究目标。