Intelligent Surface Finishing Using Compliant Robots

Abstract

Surface finishing tasks such as grinding or sanding are tedious tasks for highly skilled craftsmen and involve health risks for humans. The automation of the process by a robot can make a major contribution and also offers economic advantages. It is challenging, however, to deal with complex geometries: due to high curvatures, different areas of the surface cannot be optimally reached by a simple strategy using a tool with relatively large and flat finishing disk. In this thesis, intelligent planning and execution methods are investigated. The approach was evaluated by a practical experiment on the DLR SARA robot and in simulation. High coverage was achieved for a complex freeform geometry.