A new method for identifying ultrasonic echo signal features using adaptive second generation wavelet

To identify the features of a hydraulic system's ultrasonic echo signal corrupted by noise, we develop a new method for constructing adaptive second generation wavelet. This new method is explained in sections 1, 2 and 3. Their core consists of: (1) based on the second generation wavelet, we use the size of kurtosis of ultrasonic echo signal as objective function to select its prediction operator and updating operator, each of which can adaptively match its features at each level; thus, the ultrasonic echo signal is predicted and updated; (2) according to the level of noise of the ultrasonic echo signal decomposed by using the adaptive second generation wavelet at each level, we select the de-noising threshold value to de-noise and recover the ultrasonic echo signal and extract its features in time domain. Section 4 and 5 apply our method to the feature extraction of ultrasonic echo signal; the results, given in Figs. 1 through 7, and their analysis show preliminarily that our method for constructing the adaptive second generation wavelet can effectively extract the features of ultrasonic echo signal by identifying its arrival time point and the interval between two time points, indicating that our method outperforms the traditional wavelet transform methods.