Global Linearized Sparse Prediction and Adaptive Dead Zone Compensation for a Piezoelectric Actuator

Area of Science:

• Robotics and Control Systems
• Materials Science and Engineering
• Mechatronics

Background:

• Piezoelectric actuators (PEAs) are crucial for high-precision motion but suffer from nonlinearities like velocity dead zones and hysteresis.
• These nonlinearities limit the performance and accuracy of PEA systems in demanding applications.
• Addressing these limitations is essential for advancing precision engineering and robotics.

Purpose of the Study:

• To introduce a novel control framework for dual-mode standing wave PEAs to overcome inherent nonlinearities.
• To specifically address and compensate for the velocity dead zone in PEA systems.
• To enhance the precision and robustness of PEA velocity regulation.

Main Methods:

• Integration of a Global Linearized Sparse Prediction (GLSP) model using Koopman operator theory for linearization.
• Implementation of an Adaptive Kalman Observer-based Model Predictive Control (AKOBMPC) strategy for disturbance estimation and control optimization.
• Utilizing a deep learning-based structured pruning mechanism within the GLSP model for computational efficiency.

Main Results:

• The GLSP model accurately predicted PEA behavior under varying loads and excitation frequencies.
• The AKOBMPC controller effectively suppressed the velocity dead zone, achieving tracking errors within ±0.35 mm/s for trapezoidal and ±0.50 mm/s for sinusoidal references.
• Demonstrated superior performance compared to conventional control methods in high-precision velocity regulation.

Conclusions:

• The proposed GLSP and AKOBMPC framework offers a robust solution for precise velocity control in PEA systems.
• This advancement contributes to the development of next-generation precision actuators with improved performance.
• The study highlights the potential of Koopman operator theory and adaptive control for complex nonlinear systems.