Compressive channel estimation techniques for narrowband or wideband communications systems

自适应稀疏信道估计技术已经在不同的无线通信系统中得到了研究。1)基于D尺度信道模型，提出了两种算法：稀疏归一化最小四乘(NLMF)算法和混合最小二乘/四乘(LMS/F)算法。这两种算法可以充分利用D尺度信道的稀疏性。2)针对非高斯噪声模型，提出了一种稳定的稀疏符号最小均方(SLMS)算法来抑制脉冲噪声，充分利用信道的稀疏性。给出了典型的仿真结果，验证了所提算法的有效性。此外，本项目还对该算法的正则化参数选择进行了研究。参数的选取可以改善算法的估计性能，同时加快算法的收敛速度。

Adaptive sparse channel estimation techniques have been studied in different wireless communication systems. To reduce computational complexity and to achieve the robustness in different noise environment, in this project, our research results are summarized as follows:1) Based on the D-scale channel model, we proposed two algorithms: sparse normalized least mean fourth (NLMF) algorithm and mixed least square/fourth (LMS/F) algorithm. The two proposed algorithms can take advantage of the channel sparsity in the D-scale channel. Computer simulation results have been provided to confirm the effectiveness of the proposed algorithms.2) Based on the non-Gaussian noise model, we proposed a kinds of stable sparse sign least mean square (SLMS) algorithms to mitigate impulsive noise and to exploit channel sparsity. Representative simulations have been given to validate the proposed algorithms. In addition, regularization parameter selection for the proposed algorithm has been investigated in this project. The selected parameter can improve the estimation performance while accelerate the convergence speed for the proposed algorithms.