Search results

Due to the advantages of fast response, high positioning precision and large stiffness, the piezoelectric-actuated nanopositioning stage is widely used in the micro/nanomachining fields. However, due to the inherent nonlinear hysteresis of the piezoelectric-actuator, the positioning accuracy of nanopositioning stage is greatly degraded. Besides, the nanopositioning stage is always performed with repetitive trajectories as the reference signals in applications, which makes the hysteresis behavior periodic. To this end, an adaptive resonance suppression iterative learning control (ARS-ILC) is proposed to address the hysteresis effect. With this effort, the positioning accuracy of the nanopositioning stage is improved.

The hysteresis behavior is identified by the Prandtl–Ishlinskii model. By establishing a convergence function, it is demonstrated that the learnable band of ILC is restricted by the lightly damping resonance of nanopositioning stage. Then, an adaptive notch filter (ANF) with constrained poles and zeros is adopted to suppress the resonant peak. Finally, online stability supervision (OSS) is used to ensure that the estimated frequency converges to the resonant frequency.

A series of experiments were carried out in the nanopositioning stage, and the results validated that the OSS is available to ensure the convergence of the ANF. Furthermore, the learnable band was extended via ARS-ILC; thus, the hysteresis behavior of nanopositioning stage has been canceled.

Due to high accuracy and easy implementation, the ARS-ILC can be used in not only nanopositioning stage control but other fabrication process control with repetitive motion.