Adaptive wavelet neural network control with hysteresis estimation for piezo-positioning mechanism

Area of Science:

• Mechatronics
• Control Systems Engineering
• Robotics

Background:

• Piezo-positioning mechanisms suffer from performance degradation due to inherent hysteresis.
• Accurate modeling and control of hysteresis are critical for precise positioning.

Purpose of the Study:

• To develop an adaptive wavelet neural network (AWNN) control strategy for piezo-positioning mechanisms.
• To address the challenge of hysteresis and improve overall control performance.

Main Methods:

• A novel hysteretic model integrating a modified hysteresis friction force function was developed.
• An AWNN controller utilizing a wavelet neural network (WNN) for function approximation was designed.
• A robust compensator was incorporated to handle uncertainties and disturbances.
• Adaptive learning algorithms based on Lyapunov stability theorem were derived for online parameter tuning.

Main Results:

• The proposed AWNN controller effectively approximated unknown dynamics and compensated for hysteresis.
• Experimental results demonstrated improved command tracking performance.
• The control system exhibited robustness against external load disturbances.

Conclusions:

• The developed AWNN control with hysteresis estimation significantly enhances the performance of piezo-positioning systems.
• The proposed approach offers a robust solution for precise control in the presence of hysteresis and uncertainties.