基于压缩感知的MIMO-OFDM系统稀疏信道估计方法

Compressive Sensing-Based Sparse Channel Estimation Method for MIMO-OFDM Systems

在多输入多输出正交频分复用(MIMO-OFDM)系统中,信号经过频率选择性衰落的信道后,在接收端需要进行均衡和相干信号的检测,故准确的信道估计量必不可少。传统的信道估计方法均基于信道抽头是密集型的假设,利用线性重构算法,如最小二乘(LS)或最小均方误差(MMSE)等,可以达到Cramer-Rao下界(CRLB)。然而,通过物理信道测量发现,在实际通信系统中,宽带信道抽头分布通常表现出稀疏特性。通过充分利用信道的稀疏特性,该文将压缩感知中的CoSaMP重构算法应用于MIMO-OFDM系统的稀疏多径信道估计。在达到与传统的信道估计方法相同性能的前提下,基于CoSaMP的信道估计方法以非常小的计算复杂度为代价,大大减少了导频信号开销,从而提高了频谱资源利用率。

Channel equalization and coherent detection require accurate channel state information （CSI） at the receiver for multiple-input multiple-output orthogonal frequency division multiplexing （MIMO-OFDM） systems. The conventional linear recovery methods, such as least squares （LS） and minimum mean square error （MMSE）, are widely adapted in channel estimation under the assumption of rich multipath. However, numerous physical measurements have verified that the practical multipath channels tend to exhibit sparse structures. In this paper, exploiting the channel sparsity, we propose a compressive sensing-based CoSaMP recovery algorithm for MIMO.OFDM sparse channel estimation. Simulations show that the compressive sensing estimation method can obtain the accurate CSI with fewer pilots than conventional linear estimation for MIMO-OFDM systems at the cost of less computational complexity. The proposed method can greatly improve the spectrum efficiency for MIMO-OFDM systems.