Modular Approach to Video Designation of Manipulation Targets for Manipulators

摘要

This paper describes a method for enhancing semi-autonomous tele- operation of highly dexterous manipulators. The presented process enables the operator of a mobile manipulator to select an object from a remotely viewed video feed transmitted from a UGV and watch the manipulator autonomously moves its end effector to a specified offset from the designated object. The presented method is robust to sensor noise and is constructed in a modular manner that does not rely on a specific 3D sensor, manipulator, camera, or number of cameras. By grossly estimating the workspace of the manipulator and scanning the general area of interest, an octree is created in which a ray is cast into to determine the point of interest. The system is designed specifically with future Explosive Ordnance Disposal (EOD) robots in mind which will possess manipulators with high degrees of freedom.