基于LMD频域近似熵频谱感知算法研究

LMD Frequency Domain Approximate Entropy-based Spectrum Sensing Algorithm

针对现有频域近似熵频谱感知技术在低信噪比条件下抗噪声性能和检测性能有待提升的问题,提出了一种基于LMD频域近似熵的频谱感知算法。(1)算法筛选出3个PF分量累加求和,使得算法提取局部调频包络特征信息得到最优,进一步排除噪声不确定度的影响。(2)算法对累加PF分量进行频域变换后求其近似熵,增强算法对频域信息的嗅探能力,提升算法检测性能。Monte Carlo仿真结果表明,在噪声不确定度为0dB,采样点数为8 000的情况下,当信噪比大于-19 d B时,可以对2ASK信号达到100%的检测概率,与现有频域近似熵算法相比,检测性能约有17 d B的提升。

The existing spectrum sensing methods process two shortcomings, such as poor anti- noise performance and detection performance under the condition of low Signal-to-Noise Radio （SNR） simulation, a LMD frequency spectrum sensing algorithm of approximate entropy is proposed to overcome these problems. Firsty, The algorithm which selects three PF components added up to extract the optimal local Frequency and envelope feature information, can further eliminate the influence of the noise uncertainty. Secondly, The algorithm calculates the approximate entropy of the cumulative component after it is Fourier transformed to promote the algorithm＇s sniffing in frequency domain information and improve the performance of detection. The Monte Carlo simulation results show that compared with time domain approximate entropy algorithm, the detection performance of 2ASK signal has been improved about 17 dB, under the noise power uncertainty of 0 db, sampling numbers of 8 000. At the same time, the detection probability can achieve of 100 % when the SNR is higher than -19 dB.