摘要
提出了一种基于广义解调和奇异值分解的时频表示增强算法。该算法利用估计的相位函数对信号进行广义解调,通过奇异值分解对得到的解调信号时频谱进行降噪处理,计算降噪后时频谱的逆变换并进行广义逆解调,重新计算时频谱获得信号增强的时频表示。仿真结果表明,该算法与已有算法相比,时频平面的随机噪声几乎被完全抑制,是一种有效的时频表示增强方法。
A time-frequency representation enhancing algorithm is proposed based on generalized demodulation and singular value decomposition (SVD). The signal is generally demodulated with the estimated phase function, and the time-frequency spectrum of the demodulated sig- nal is enhanced with SVD. Then, the inverse time-frequency transform and the inverse demodulation are calculated. Finally, the enhanced time-frequency representation is obtained after the time-frequency spectrum of the noise-reduced is recalculated. Simulation results show that the random noise is almost completely suppressed on the time-frequency plane, and the proposed algorithm is an effective way to enhance the time-frequency representation compared with the other algorithms.
出处
《数据采集与处理》
CSCD
北大核心
2010年第4期419-424,共6页
Journal of Data Acquisition and Processing
基金
国家高技术研究发展计划("八六三"计划)(2006AA09A102)资助项目
国家自然科学基金重点(41070424)资助项目
关键词
广义解调
奇异值分解
时频表示
信号增强
generalized demodulation
singular value decomposition (SVD)
time-frequency representation
signal enhancing