传统的频谱感知能量检测易受噪声方差不确定性的影响,存在"信噪比墙"效应。频谱感知特征值检测跟能量检测一样,不需要信号任何先验信息,并且能在低信噪比下取得较好的检测性能。经典的特征值检测有最大最小特征值(maximum-min...传统的频谱感知能量检测易受噪声方差不确定性的影响,存在"信噪比墙"效应。频谱感知特征值检测跟能量检测一样,不需要信号任何先验信息,并且能在低信噪比下取得较好的检测性能。经典的特征值检测有最大最小特征值(maximum-minimum eigenvalues,MME)之比算法,最大最小特征值之差(maximum-minimum eigenvalues difference,DMM)算法等。这些算法只利用特征值的一阶统计量,不能充分反映全部特征值的统计特征。利用特征值二阶统计量提出一种基于特征值方差的频谱感知算法,选取能反映特征值整体波动的方差当作观测统计量,并利用矩阵迹的性质推导出该算法的理论门限。仿真证明:当噪声方差不确定性等于0 d B时,该算法的检测性始终优于MME算法。当噪声方差不确定性等于0. 2 d B时,能量检测(energy detection,ED)算法检测概率急剧下降,而特征值方差(eigenvalue variance,EV)算法检测概率仅下降10%左右,并且当信噪比(signal noise ratio,SNR)大于-17 d B时,EV算法的检测概率优于ED算法和MME算法。展开更多
为探索神经反馈训练在提升射击表现方面的应用效果和训练过程中的无应答者特性,开展一项用于提升射击表现的神经反馈训练(neurofeedback training for sport performance,SP-NFT)实验研究,招募20名受试者,进行2周4次的“巅峰”范式SP-N...为探索神经反馈训练在提升射击表现方面的应用效果和训练过程中的无应答者特性,开展一项用于提升射击表现的神经反馈训练(neurofeedback training for sport performance,SP-NFT)实验研究,招募20名受试者,进行2周4次的“巅峰”范式SP-NFT,采集受试者前、后测隐显目标射击表现和相关脑电(electroencephalograph,EEG)数据,检验SP-NFT对射击表现的提升效果、静息态EEG特征、SP-NFT期间正常组和无应答组EEG特性变化情况。结果表明:受试者后测射击成绩显著高于前测(P<0.01),静息态theta频带功率显著降低(P<0.01);相对正常受试者,无应答者在SP-NFT期间的努力程度更高,theta频段功率和SMR功率的变化程度更低,SP-NFT能够有效提升受试者射击表现,进一步揭示了无应答者的相关生理机制。研究成果为用于提升射击表现的SP-NFT技术进一步发展提供理论支撑和实验证据。展开更多
According to the problem of cognitive ultra wide-band （UWB） spectrum sensing, a novel UWB pulse signal detection algorithm based on cumulative sum （CUSUM） test is proposed in this paper. Based on the analysis of t...According to the problem of cognitive ultra wide-band （UWB） spectrum sensing, a novel UWB pulse signal detection algorithm based on cumulative sum （CUSUM） test is proposed in this paper. Based on the analysis of the existing spectrum sensing schemes for cognitive UWB system, some obvious facts are obtained that it is difficult to detect UWB pulse signal with conventional spectrum sensing schemes, due to its low average signal to noise ratio （SNR）, large bandwidth, and low duty ratio. In this paper the detection algorithm of signal distribution change, which is application of CUSUM test, is considered to be applied to cognitive UWB spectrum sensing. But CUSUM test request that the pre-change and the post-change distributions are i.i.d, which cannot be satisfied in the detection process of UWB pulse signal. Since there are two time domain descriptions on UWB pulse signal, namely one contains only noise and the other one contains pulse signal plus noise, the existing detection algorithm of signal distribution change cannot be directly applied to detect UWB pulse signal. Hence the uniform probability density function expression of UWB pulse signal is first deduced, then CUSUM test is applied to cognitive UWB spectrum sensing. The proposed algorithm is a time sequential detection algorithm, with low complexity and minimal detection delay, which is suitable to detect the low duty ratio signal. Its performance is evaluated through theoretical analysis and numerical simulations. It is shown that this algorithm outperforms the conventional energy detection algorithm and conquers SNR wall phenomenon.展开更多
文摘传统的频谱感知能量检测易受噪声方差不确定性的影响,存在"信噪比墙"效应。频谱感知特征值检测跟能量检测一样,不需要信号任何先验信息,并且能在低信噪比下取得较好的检测性能。经典的特征值检测有最大最小特征值(maximum-minimum eigenvalues,MME)之比算法,最大最小特征值之差(maximum-minimum eigenvalues difference,DMM)算法等。这些算法只利用特征值的一阶统计量,不能充分反映全部特征值的统计特征。利用特征值二阶统计量提出一种基于特征值方差的频谱感知算法,选取能反映特征值整体波动的方差当作观测统计量,并利用矩阵迹的性质推导出该算法的理论门限。仿真证明:当噪声方差不确定性等于0 d B时,该算法的检测性始终优于MME算法。当噪声方差不确定性等于0. 2 d B时,能量检测(energy detection,ED)算法检测概率急剧下降,而特征值方差(eigenvalue variance,EV)算法检测概率仅下降10%左右,并且当信噪比(signal noise ratio,SNR)大于-17 d B时,EV算法的检测概率优于ED算法和MME算法。
文摘为探索神经反馈训练在提升射击表现方面的应用效果和训练过程中的无应答者特性,开展一项用于提升射击表现的神经反馈训练(neurofeedback training for sport performance,SP-NFT)实验研究,招募20名受试者,进行2周4次的“巅峰”范式SP-NFT,采集受试者前、后测隐显目标射击表现和相关脑电(electroencephalograph,EEG)数据,检验SP-NFT对射击表现的提升效果、静息态EEG特征、SP-NFT期间正常组和无应答组EEG特性变化情况。结果表明:受试者后测射击成绩显著高于前测(P<0.01),静息态theta频带功率显著降低(P<0.01);相对正常受试者,无应答者在SP-NFT期间的努力程度更高,theta频段功率和SMR功率的变化程度更低,SP-NFT能够有效提升受试者射击表现,进一步揭示了无应答者的相关生理机制。研究成果为用于提升射击表现的SP-NFT技术进一步发展提供理论支撑和实验证据。
基金supported by the Aviation Science Fund (20095596014)
文摘According to the problem of cognitive ultra wide-band （UWB） spectrum sensing, a novel UWB pulse signal detection algorithm based on cumulative sum （CUSUM） test is proposed in this paper. Based on the analysis of the existing spectrum sensing schemes for cognitive UWB system, some obvious facts are obtained that it is difficult to detect UWB pulse signal with conventional spectrum sensing schemes, due to its low average signal to noise ratio （SNR）, large bandwidth, and low duty ratio. In this paper the detection algorithm of signal distribution change, which is application of CUSUM test, is considered to be applied to cognitive UWB spectrum sensing. But CUSUM test request that the pre-change and the post-change distributions are i.i.d, which cannot be satisfied in the detection process of UWB pulse signal. Since there are two time domain descriptions on UWB pulse signal, namely one contains only noise and the other one contains pulse signal plus noise, the existing detection algorithm of signal distribution change cannot be directly applied to detect UWB pulse signal. Hence the uniform probability density function expression of UWB pulse signal is first deduced, then CUSUM test is applied to cognitive UWB spectrum sensing. The proposed algorithm is a time sequential detection algorithm, with low complexity and minimal detection delay, which is suitable to detect the low duty ratio signal. Its performance is evaluated through theoretical analysis and numerical simulations. It is shown that this algorithm outperforms the conventional energy detection algorithm and conquers SNR wall phenomenon.
