An adaptive weighted least square support vector regression for hysteresis in piezoelectric actuators

To overcome the low positioning accuracy of piezoelectric actuators (PZAs) caused by the hysteresis nonlinearity, this paper proposes an adaptive weighted least squares support vector regression (AWLSSVR) to model the rate-dependent hysteresis of PZA. Firstly, the AWLSSVR hyperparameters are optimized by using particle swarm optimization. Then an adaptive weighting strategy is proposed to eliminate the effects of noises in the training dataset and reduce the sample size at the same time. Finally, the proposed approach is applied to predict the hysteresis of PZA. The results show that the proposed method is more accurate than other versions of least squares support vector regression for training samples with noises, and meanwhile reduces the sample size and speeds up calculation.