Next-generation highly precision monitoring system for fishing boat based on waveform separation

提出了一种利用抛物面声反射式传声器和被监测设备产生的合成声分离信号处理技术对动态多个设备进行状态监测的新方法。通过估计驱动频率的强度差，对几种设备的驱动频率分量进行了识别。利用抛物面传声器在特定方向上具有较高灵敏度的方向性来估计驱动频率分量的强度差。其次，通过对信号的频域幅值和相位值进行信号处理，提取出区分后的独立驱动频率分量。最后，通过逆FFT将提取的单个驱动频率分量转换成时间序列。可以使用比例方法从多个设施中提取由状态监测设施产生的时间序列数据。

This research proposes a novel condition monitoring method for dynamic plural faciities using a parabolic sound reflector type microphone and separation signal processing technique about composed sound generated by the monitored facilities. Discrimination of driving frequency components of several facilities has been performed by estimating intensity difference of the driving frequency. The intensity differences of the driving frequency components are estimated by the directivety of an applied parabolic microphone which has high sensitivity in the specific direction. In the next place, the discriminated indivudual driving frequency components are extracted by the signal processing operation about its amplitude and phase values in frequency domain. Finally, extracted individual driving frequency components convert into time series of by invserse FFT. Time series data generated by the condition monitored facility from plural facilities can be extracted using the proporsal method.