Abstract:Abstract: In order to reduce the amount of data of non-stationarity and nonlinearity signal sent throughout the sensor network, a grey morlet wavelet kernel partial least squares (GMWKPLS) model is proposed. In this model, grey prediction theory is firstly introduced into kernel partial least squares (KPLS). Then the input-output data are mapped to a nonlinear higher dimensional feature space by morelt kernel transformation. Finally, a prediction fusion model is constructed by linear partial least squares. Moreover, moving widow method is utilized to update samples continuously in this dynamical prediction model. The model is validated with vibration signals of gear tooth breakage in the speed of rising. The results show that the model can execute dynamic multi-step prediction, and it has high precision prediction. Thus, it can observably reduce the amount of sending data throughout the sensor network and save energy. Comparing with grey RBF kernel partial least squares (GRBFKPLS) and RBF kernel partial least squares (RBFKPLS), GMWKPLS is best in prediction performance, and the prediction errors are around ±0.4%.