【摘要】 在正交频分复用（OFDM）系统信道估计中，以阈值为基础的离散傅里叶变换（DFT）去噪算法中阈值的选取非常关键，阈值过大会滤除有用信号，过小则无法滤除噪声。针对此问题，本文提出一种改进的基于二次阈值的DFT去噪算法，即利用循环前缀（CP）外噪声样本来设置第一次阈值以滤除CP内噪声样点，然后在此基础上结合CP内外信号样本设置第二次阈值门限，以尽量减小第一次阈值设置过大或过小对信号造成的干扰。仿真结果表明，与现有DFT阈值门限算法相比，本文算法估计性能更好，在信噪比≥15 dB时，系统误码率降低了50%。在系统误码率为10^-1或更低时，本文算法至少有3 dB的性能提升，且能获取较为准确的信道信息，进一步减小了系统均方误差和误码率。更多还原

【Abstract】 In Orthogonal Frequency Division Multiplexing（OFDM） system channel estimation, the selection of threshold value in the threshold-based Discrete Fourier Transform（DFT） denoising algorithm is very critical, and the threshold value is too large to filter out the useful signal, and too small to filter out the noise. To address this problem, this paper proposes an improved DFT denoising algorithm based on the second threshold, i.e., using the noise samples outside the Cycle Prefix（CP） to set the first threshold to filter out the noise samples inside the CP, and then set the second threshold threshold based on this combination of inside and outside the CP signal samples, in order to minimize the interference of the signal caused by the first threshold being set too large or too small. Simulation results show that the proposed algorithm has better estimation performance than the existing DFT threshold thresholding algorithms, and the system BER is reduced by 50% when the SNR is 15 dB or higher, and the proposed algorithm has at least 3 dB performance improvement when the system BER is 10^-1 or lower, and it can obtain more accurate channel information, which further reduces the system mean-square error and BER.更多还原