Abstract: In combination of CS theory, the compressing sampling matching pursuit algorithm for Multiple-Input Multiple-Output Orthogonal Frequency Division Multiplexing（MIMO_OFDM） system channel estimation requires the signal sparsity as a priori information, while in actual situation the sparsity is difficult to obtain, for this question it proposes a signal sparsity adaptive Compressive Modifying Sampling Matching Pursuit algorithm（CoMSaMP）. The algorithm adopts the atomic weak selection criteria with theoretical support as a pre-selection scheme, and sets the first clipping threshold to reduce the algorithm extra iteration, then reduces the computational complexity, the improve of crop mode on channel estimation ensures the improvement of the reconstruction accuracy, and ultimately realizes adaptive recovery on MIMO-OFDM sparse channel estimation . Simulation results show that, compared with the original algorithm, under the same SNR conditions, the CoMSaMP  algorithm has better performance on channel estimation, improves the spectral efficiency, reduces the complexity. When the sparsity level is high, the proposed algorithm has the better performance than the CoSaMP algorithm on anti-interference ability.

Key words: compressed sensing, Orthogonal Frequency Division Multiplexing（OFDM）, sparse channel estimation, Compressive Sampling Matching Pursuit（CoSaMP）

摘要： 结合压缩感知理论（CS），针对压缩采样匹配追踪算法在多输入多输出正交频分复用（MIMO_OFDM）系统信道估计应用中需要利用信号稀疏度的先验条件，而实际中稀疏度又难获得的情况，提出一种信号稀疏度自适应的压缩采样改进匹配追踪算法（CoMSaMP）。该算法采用具有理论支撑的原子弱选择标准作为预选方案，并设置首次裁剪阈值来减少算法多余的迭代，降低算法在信道估计中的复杂度，裁剪方式的改进保证了重构精度的提高，最终实现MIMO-OFDM稀疏信道估计中信号的稀疏度自适应。仿真结果表明：与原算法相比，该算法在同等信噪比条件下具有更优的信道估计性能，从而提高了频谱利用率，同时降低了复杂度，在稀疏度较高时，提出的算法具有更好的对噪声的抗干扰能力。

关键词: 压缩感知, 正交频分复用, 稀疏信道估计, 压缩采样匹配追踪