Precise and Robust Large-Shape Formation Using Uncalibrated Vision for a Virtual Mold

Consider four flat-surface regions each of which has been placed, with varying positions and orientations, within the workspace of a robot, as suggested in Figure 1. Assume that the robot is kinematically able to reach each surface in the sense of delivering a cutting and an abrading tool to the poses needed in order to remove material from a blank to the point of recovering each of these surfaces. Fig. 1. The four flat surfaces which are to be duplicated via robot machining and abrasive action – precisely in their respective locations. We pose the task as one of recovering to within a maximum error of 0.1mm, each of the four surfaces of Fig. 1 using the robot with end-of-arm machining and abrading equipment.