水声多输入多输出信道的分布式压缩感知估计

多输入多输出技术通过采用多个阵元进行多发多收空间复用信道可在极其有限的通信带宽下实现高速水声通信,但由于同时存在通道间干扰和多径干扰,水声MIMO信道估计变得困难。提出利用MIMO水声信道多径稀疏结构存在的相关性,在经典联合稀疏模型的基础上对MIMO观测矩阵进行重组,从而建立基于分布式压缩感知的单载波水声MIMO通信信道联合稀疏模型;同时,针对信道响应中具有相同多径位置的稀疏部分和特有稀疏部分设计区分性正交匹配追踪算法进行联合重构,进一步抑制通道间干扰的影响。最后通过仿真和海上实验进行本方法有效性的验证,实现16 kbps的MIMO水声通信。通过算法推导、仿真和实验可得到结论:利用MIMO水声信道多径相关性进行分布式压缩感知估计可提高估计性能。 

Distributed compressed sensing estimation of underwater acoustic multiple-input-multiple-output channels

MIMO (Multiple Input Multiple Output) technology provides a potential solution for high data rate under- water acoustic communications under limited bandwidth. However, simultaneous presence of multipath and co-channel interference (Co-channel interference, CoI) poses significant difficulty to estimation of acoustic MIMO channels with the conventional method such as classic algorithms or compressed sensing (CS) algorithms. To exploit the spatial correlation feature of the acoustic MIMO channels, the distributed compressed sensing (DCS) acoustic MIMO channel estimation model based on re-organizing of the MIMO measurement matrix is proposed. By discriminatively estimating the sparse components with the same time delay and those with different time delay, a novel DOMP (Discriminative Orthogonal Matching Pursuit) algorithm is designed to facilitate enhanced estimation of multipath components, as well as alleviation of the CoI. Numerical simulations as well as sea trial experiments are provided to demonstrate the superior performance of the proposed method.