An Application of Tool Insertion Using Model Based Vision

Abstract

Model-based vision is shown to be a viable solution to a tool insertion problem for the nuclear servicing industry. A complete tool insertion system requires: representation of the model, extraction of model parameters from sensor data, quantification of sensor error, transformation of sensor error to model error, sensor calibration techniques, and motion planning for sensor drop out recovery. Each of these topics is considered as a component technology and related as an aggregate system in a technical insertion scenario. While each of these technologies is required in a complete system, the specific technique is application dependent. The requisite technologies are discussed in the context of the general problem, the insertion of a four-pronged tool into four specific holes in a plane, using a vision-corrected CAD model.