Smart sensors for autonomous robotic panel assembly

Abstract

Manufacturing of a typical aluminium airframe panel involves the assembly of a couple of stringers and several dozens of clips. The stringers have to be mounted to the skin by riveting, while the clips are glued by using a fast hardening sealant. Though there are stringers and clips of identical geometry, there is too much part variation to make automation an easy task, and since tolerances are introduced by sagging manual work is still state of the art in order to cope with this. The paper is about the implementation of a computer vision sensor system that is capable of identifying individual clips by their CAD-drill hole pattern, determine the clips position and allow a robot to correctly pick and place them regardless of positional tolerances. In addition to that, a high accuracy drill hole detector that enables to correctly pick and place stringers and compensates the sagging effects, has been implemented. To ensure the quality of the detection processes the position of the individual parts have been investigated by a structured-light 3D-Camera. Within the paper the involved measurement technologies and the robot calibration are focused. An overview of the entire process chain is also given as well as an outlook to future applications in related areas.