A method of feature extraction in weak current signals

To extract weak current signal features in the spindle current of machine tools, which are often buried in noises, a signal processing method with modulated stochastic resonance and D-J threshold theory is proposed. Modulated stochastic resonance is adopted to amplify the features of weak current signals, and then D-J threshold theory is chosen to make noise estimation for extracting the useful signals. The amplitude is estimated by the stochastic resonance. The simulated and experimental results show that this method enables to accurately identify the weak feature frequencies and estimate the amplitudes of weak current signals properly.