Abstract: Because servo current of AC servo motor is subject to many disturbances, from which it is difficult to obtain useful information directly, the servo current coming from the controller of AC servo motor is measured by means of Hall current sensor, and then wavelet transformation theory is applied to extract useful information reflecting the dynamics of friction from this servo current. In this process, multi-resolution analysis is performed upon the servo current, which means the servo current signal is divided into several levels of details and a smooth part, from which the feature of information from static friction to dynamic friction is detected. Compared with the output signal of displacement, it is shown that information influenced by friction can indeed be extracted from the servo current. It is also found that the sudden change in the servo current always appears ahead of macroscopic movement of the worktable, which needs further study.

Key words: Feature extraction, Friction, Servo current, Wavelet transform