Kinematic Calibration of a 3DOF Compliant Nanopositioning Stage Based on Visual Measurement

This paper demonstrates a vision-based kinematic calibration method for a 3DOF compliant nanopositioning stage (CNPS). First, the nanopositioner's simplified linear elastic kinematics model is established based on the pseudo-rigid body model. The vision-based multi-DOF pose measurement scheme that can accurately measure the milliradian rotation angle is given, including the measurement system and principle. To accurately calibrate the kinematic model, the motion transfer Jacobian matrix is split into a translation part and rotation part, and different pose data and direct linear transformation are used to solve them, respectively. Finally, the calibrated kinematic parameters are utilized for CNPS positioning control. The results show that the rotation error of CNPS is reduced from 0.0232° to 0.00380°, and the translation error is reduced from 0.3754μ
$m$
to 0.0612μm.