MIMO水声信道联合稀疏模型与分布式压缩感知估计方法研究

MIMO technology offers an effective solution for the high data rate underwater acoustic (UWA) communication. However,multipath, cochannel interference (CoI) as well as time variations of MIMO underwater acoustic channels pose significant difficulty for MIMO channel estimation. The present project proposes to exploit the joint spatial sparsity of MIMO UWA channels to address this problem under the framework of distributed compressed sensing (DCS). Firstly, spatial correlation of MIMO UWA channels will be analyzed to setup and verify the joint spatial sparse model to enable the adoption of distribution compressed sensing (DCS) reconstruction. Meanwhile, the joint observation matrix is re-organized to decompose the MIMO channels into the common sparse components and the different sparse component to enable joint estimation as well as mitigation of the CoI. The common sparse components and the different sparse components are estimated by stages with joint sparse reconstruction method and the fast sparse recovery method respectively. The total MIMO channel response is obtained by summing of both components. Finally, impact of the complicated spatial-temporal features of UWA channels on the joint sparse model and DCS estimation method is investigated. The innovative progresses achieved by this project on DCS estimation of MIMO UWA channels have the potential of being employed in the R&D of high performance UWA MIMO commmunication and detection systems.

MIMO技术为在带宽严重受限的水声信道中实现高速通信提供了一种有效途径。但复杂的多径、通道间干扰及时变特性对MIMO水声信道估计造成严重困难，本项目拟利用MIMO水声信道的联合稀疏特性进行分布式压缩感知框架下的高效估计。.首先，分析水声多发多收信号多径传输导致MIMO信道具有的空间稀疏相关特性，建立MIMO水声信道的联合稀疏模型并进行有效性验证，在该模型下MIMO信道具有公共稀疏支撑集，从而可提高稀疏重构概率；考虑通道间干扰的影响，研究重组联合测量矩阵将MIMO信道分解为共同稀疏部分和差异稀疏部分，分别利用联合稀疏重构及快速稀疏恢复算法对两部分进行分步骤估计，并叠加后得到估计结果；分析水声信道复杂空时特性对稀疏相关特性及联合估计性能的影响。.项目在MIMO水声信道联合稀疏估计上的创新和突破对高性能MIMO水声通信及探测系统研究有积极意义。

MIMO通信提供了在恶劣水声信道条件下实现高速率通信的有效手段，但是，与单发射通信系统相比MIMO通信系统不但存在多径、时变等水声信道恶劣传输特性，而且存在不同发射阵元在同一接收阵元所造成的同道干扰，导致传统MIMO信道估计方法性能不可避免的下降，对高性能MIMO水声通信系统的研究和应用造成极大的挑战。.分布式压缩感知（Distributed Compressed Sensing, DCS）理论是在压缩感知的基础上针对多个信号稀疏性具有相关的现象而提出，通过利用多个信号的共同稀疏性进行联合重构可进一步提高稀疏重构性能。项目在分布式压缩感知的框架下将MIMO水声信道估计转换成空、时联合稀疏恢复问题，通过对接收到的水声通信信号各数据块间的多径结构存在的较强空、时相关性进行高效稀疏估计从而提高估计性能；同时，项目提出了带干扰抵消模块的信道估计取得MIMO信道均衡器结构以提高MIMO通信接收机与MIMO信道的适配性从而改善通信性能。项目通过数值仿真和厦门海域海试实验结果充分验证了项目所提MIMO信道联合稀疏估计方法及MIMO通信接收机方案的有效性，试验数据表明采用项目技术方案的海试实验MIMO通信接收机可获得约1dB的输出信噪比增益。项目研究结果表明：通过对MIMO水声信道存在的空、时稀疏相关特性进行充分利用，结合干扰抵消接收机结构，可获得通信性能改善。项目相关研究成果在包括声学领域国际顶级期刊（美国声学学会会刊）在内的学术期刊发表了15篇学术论文；同时，部分研究成果已应用于项目团队系列水声通信设备研发。.项目在MIMO信道模型、高效估计方法及接收机结构方面的研究成果可推广应用于高性能水下通信系统研发及工程化应用，对提高海洋工程、环境保护、国防安全等领域高速率水下数据传输性能有积极意义。