Comparison of Pose Error Compensation for Focal Plane Pose Test Platform Using GRNN and CART

摘要

The surface accuracy of the telescope focal plate plays a key role in high-precision astronomical observations. The 6-DOF parallel Focal Plane Pose Test Platform (FPPTP) is used to measure the deformation and surface accuracy of the focal plate in different space pose, and precise pose adjustment is an important indicator of the platform's performance. But the factors affecting the pose error of the platform are complex and difficult to describe accurately with mathematical model. Comparison of pose error compensation for the focal plate in different space pose using Generalized Regression Neural Network (GRNN) and Classification Regression Tree (CART) is studied in this paper.