Pose Sensing and Servo Control of the Compliant Nanopositioners Based on Microscopic Vision

This article presents new pose sensing and servo control techniques for the compliant nanopositioners (CNPs) based on optical microscopic vision. A visual pose tracking algorithm (VPTA) and a visual servo positioning scheme (VSPS) that both utilize iterative template matching are presented. In the VPTA, to realize pose sensing of the CNPs with high performance, an improved Gaussian- Newton optimization method combined with an adaptive penalty strategy is developed. In the VSPS, to realize robust and flexible control of the CNPs, a velocity controller that directly uses the gray value of the template to control the CNP is designed. Simulations and experiments are performed to demonstrate the performance of the proposed method. Results show that the VPTA can achieve pose tracking of the three-degree-of-freedom (x, y, 0) CNPs at a frame rate of hundred hertz, and the dynamic tracking errors are smaller than 100 nm, 160 nm, and 40 μrad in the x ^- , y ^- , and 0-axes, respectively. Moreover, by using the proposed VSPS, task-based nanopositioning can be easily realized without extracting features of the object, and the obtained stable positioning accuracies are better than 30 nm, 33 nm, and 3 μrad in the x ^- , y ^- , and 0-axes, respectively.