A Neural-Network-Based Controller for Piezoelectric-Actuated Stick-Slip Devices

Piezoelectric-actuated stick-slip device (PASSD) is a highly promising equipment that is composed of one end-effector, one piezoelectric actuator (PEA) and one driving object adhered to the PEA. Since the endeffector can slip on the surface of the driving object, the PASSD is capable of realizing the macrolevel motion with the microlevel precision. Due to the following two reasons: the complicated relative motion between the end-effector and the driving object, and the inherent hysteresis nonlinearity in the PEA, the ultraprecision displacement control of the end-effector of PASSDs raises a real challenge, which is rarely reported in the literature. Toward solving this challenge, a neural-network-based controller is proposed in this paper. First, a neural-network-based model is proposed to capture the relative motion between the end-effector and the driving object. Second, a neural-network-based inversion model is developed to online calculate the desired position of the PEA under the predesigned reference of the end-effector. Third, a dynamic linearized neural-network-based model predictive control method, which can effectively handle the hysteresis nonlinearity, is employed to implement the displacement control of the PEA, which finally results in an overall high-precision controller of the end-effector. Finally, a PASSD prototype has been implemented and tested through experimental studies to demonstrate the effectiveness of the proposed approach.