摘要
针对舰船水压场和海浪水压场在频率成分上的差别,提出了小波能谱估计和支持向量机多步预测检测的舰船水压场检测方法。使用滑动加窗信号的低频小波能谱估计值作为新的检测序列,提高了被检测信号的信噪比,使用支持向量机对其进行多步预测滤波检测,并使用实测的数据进行了验证。在海浪水压场为5级海况的情况下(信噪比约为-23 dB),准确可靠的检测到了目标信号,较大的提高了舰船水压信号的检测水平。
In light of the difference between ship hydrodynamic pressure and wave pressure, a method for detecting the hydrodynamic pressure signals of ship was proposed based on wavelet power spectra estimation and support vector machine （SVM） multi-step prediction. The low frequency wavelet power spectrum of mov- ing windowing dada is used as the new detecting sequence to improve the signal- noise ratio （SNR） of detected dada, and the SVM multi-step predicting filtering is used to detect the target signal. In 5 oceanic condition （the SNR is about -23dB）, the signal can be detected accurately with good performance.
出处
《船海工程》
北大核心
2008年第3期135-138,共4页
Ship & Ocean Engineering
基金
重点实验室基金项目资助(51444060101JB1108)
关键词
小波能谱估计
支持向量机
舰船水压场
信号检测
wavelet spectra estimation
support vector machine
ship hydrodynamic pressure
signal detection
