Division of CNT placement meta-task for micro-nano robotic manipulator

The process of operating a micro-nano robotic manipulator in a scanning electron microscope (SEM) to place carbon nanotube (CNT) on the surface of a nanodevice is very difficult. Not to mention allowing the micro-nano robotic manipulator to learn its motor skills from the demonstration of the placement task. In this paper, this complex assembly task is divided into multiple simple meta-tasks, from each of which the motor skills for the placement operation can be more easily learned. We identified the AFM cantilever in each frame by template matching and calculated its position coordinates to obtain the motion trajectory of the cantilever during the whole manipulation. The sharpness of some regions of the cantilever is evaluated using the Tenengrad gradient function. And based on the specificity of the working environment of the micro-nano robotic manipulator, we established applicable division criteria, analyzed and calculated the required division parameters from the captured video of the manipulation process, and successfully divided multiple simple meta-tasks based on the information.