论文部分内容阅读
压缩感知信道估计可利用信道稀疏特性提高估计性能,但对于具有典型块稀疏分布的水声信道,经典的l_0或l_1范数无法很好地描述块稀疏特性。利用水声信道块稀疏分布规律特性提出一种能够识别块稀疏结构的块稀疏似零范数,并在稀疏恢复信道估计算法中引入块稀疏似零范数约束项,进一步推导了复数域块稀疏似零范数恢复迭代算法,该算法通过对块稀疏似零范数进行梯度下降迭代并将梯度解投影至解空间来获得水声信道的块稀疏似零范数估计。数值仿真和海上水声通信实验结果表明该算法相对经典的稀疏信道估计算法有较明显的性能改善。通过算法推导、仿真和实验可获取结论:利用水声信道的块稀疏特性进行压缩感知重构可有效提高信道估计性能。
Compressed sensing channel estimation can exploit channel sparseness to improve estimation performance, but classical l_0 or l_1 norm can not describe the block sparseness well for underwater channel with typical block sparse distribution. In this paper, we propose a block sparse zero-norm that can recognize the sparse structure of blocks based on the sparse distribution rule of underwater acoustic channel blocks, and introduce sparsity-like zero-norm constraint in the sparse recovery channel estimation algorithm. Furthermore, It resembles zero norm recovery iterative algorithm, which obtains sparse zero-norm norm of underwater acoustic channel by iteratively descending the sparse zero-norm of the block and projects the gradient solution to the solution space. The numerical simulation and experimental results of marine underwater acoustic communication show that the proposed algorithm has obvious performance improvement over the classical sparse channel estimation algorithm. Through algorithm deduction, simulation and experiment, we can get the conclusion: Compressive sensing reconfiguration using the sparse feature of underwater acoustic channel can effectively improve the channel estimation performance.