Underwater robot propeller fault feature extraction method based on thrust deviation and cross-correlation algorithm

The invention provides an underwater robot propeller fault feature extraction method based on thrust deviation and a cross-correlation algorithm, and belongs to the technical field of underwater robot fault diagnosis. According to the method, two thrust deviation curves are obtained according to longitudinal and heading degree-of-freedom related signals, and a fault occurrence moment is determined based on step point positions of the curves, so that the positioning precision of the fault occurrence moment is improved; and then, in order to strengthen fault features, designing a multi-time window Fourier transform and fusion method to respectively extract fault features of the two thrust deviation curves, solving cross-correlation fault features of the two fault features, and fusing the three fault features to obtain final fault features. On the premise of extracting a plurality of signal fault features, the method can achieve the accurate positioning of the fault occurrence time, obtains ...