A fast and accurate piezoelectric actuator modeling method based on truncated least squares support vector regression

Area of Science:

• Engineering
• Materials Science
• Control Systems

Background:

• Piezoelectric actuators (PEAs) are crucial for precision positioning.
• Modeling hysteresis in PEAs is essential for accurate control.
• Existing least squares support vector regression (LS-SVR) methods lack robustness against noise.

Purpose of the Study:

• To develop a robust hysteresis modeling technique for PEAs.
• To enhance the engineering applicability of PEA models.
• To improve the accuracy and reliability of PEA positioning.

Main Methods:

• Proposed a robust truncated least squares support vector regression (T-LSSVR).
• Implemented a truncation strategy to reduce training set redundancy and improve robustness.
• Optimized T-LSSVR parameters using particle swarm optimization and cross-optimization algorithms.

Main Results:

• The T-LSSVR method demonstrated superior accuracy and robustness compared to standard LS-SVR, especially with noisy training data.
• The proposed approach effectively reduced the required sample size.
• Enhanced computational efficiency was observed.

Conclusions:

• T-LSSVR offers a more robust and efficient solution for modeling PEA hysteresis.
• The method significantly improves the practical applicability of PEAs in precision engineering.
• This work provides a reliable approach for controlling PEAs in noisy environments.