Dynamic hysteresis modeling of piezoelectric actuator in Scanning Tunneling Microscope

Piezoelectric ceramics actuator is widely used in ultra high precision and tracking mechanism for the advantages of simple construction, high response frequency, rapid dynamic performance and excellent heavy carrying capacity. But the hysteretic nonlinear characteristic reduced the tracking precision. A modified modeling method based on dynamic recurrent neural network(DRNN) is designed in this paper to improve the tracking performance. The mechanical structure is introduced, and a Bouc-Wen model is given to express the nonlinear kinetics. The data pairs including driving voltage and corresponding displacement are regarded as the samples to train the network off-line. The weight values in DRNN are modified according to the error between the actual and desired displacement. A triangle voltage with variable amplitude is applied to validate the effectiveness of the proposed method. It is shown in the experiments that the mean tracking error is reduced from 0.38μm to 0.24μm, and the maximum error from 0.74μm to 0.42μm respectively compared with the static neural network. A more accurate model is established for the control system design in the future.