Automatic control for a miniature manipulator based on 3D vision servo of soft objects

This paper presents a new strategy for automating three-dimensional miniaturized manipulators using a visual feedback. This automation scheme can be used to handle the challenges associated with manipulation of minute soft components such as biological entities. The developed vision control system combines a depth and planar motion control using a single camera and an auto focus algorithm. This visual process was conducted by applying image segmentation and probe recognition along with depth alignment. To demonstrate the feasibility, reliability and robustness of this newly proposed strategy, extensive computer simulations were performed on images of a Zyvex® nanomanipulator obtained using a Scanning Electron Microscope (SEM). Also, a prototype manipulator with a visual feedback and an automatic controller was constructed and tested using soft elements. The results showed good accuracy and robustness of the end-effector recognition, positioning and aligning without occlusion difficulties.