基于压缩感知的稀疏多径信道估计

CS-based sparse multipath channel estimation

提出了一种基于压缩感知理论的稀疏多径信道估计方法。利用训练序列设计了一种简化的Toeplitz结构观测矩阵,证明了观测矩阵满足限制等距特性,可以作为压缩感知的观测矩阵。根据此矩阵的近似正交性特点对正则化迭代硬阈值算法进行简化,并引入精英策略提出一种归档正则化迭代硬阈值估计算法。仿真结果表明,该估计方法相对于迭代最小二乘法具有更优的性能,且提出的归档正则化迭代硬阈值算法兼具收敛速度快和稳定性高的优点。

A sparse muitipath channel estimation method based on compressed sensing （CS） is proposed. A simplified Toeplitz-structured measurement matrix is designed using the training sequence. It is proved that the measurement matrix satisfies the restricted isometric property （RIP）, which can be utilized as a measurement matrix of CS. Then, the normalized iterative hard thresholding （NIHT） algorithm is simplified by exploiting the approximate orthogonality feature of the matrix, as well the elite strategy （ES） is introduced to propose an archiving-based normalized iterative hard thresholding （ANIHT） estimation algorithm. Simulation results show that the proposed channel estimation method achieves a better performance than the traditional recursive least square （RLS） algorithm and the proposed ANIHT algorithm has the merits of high convergence speed and good stability.