An Additive Manufacturing Path Generation Method Based on CAD Models for Robot Manipulators

Abstract

Traditional extrusion based Additive Manufacturing (AM) is realized using a 3 Degrees\nof Freedom (DOF), translation only, 3D printer. Here, the printer must be larger than\nthe printed part. One way of enabling AM in large-scale is to combine AM with\nrobotics. By using a 6 DOF robot manipulator to extrude a fast-curing material, the\nworkspace of the build would be greatly expanded and it would be possible to increase\nthe flexibility of the building process itself since the structure would no longer have\nto be built from the bottom-up approach which is necessary for most existing forms\nof AM. This could possibly reduce the need for support structures to the point of\nonly relying of anchoring and stabilizing. In this thesis, a method for generating a\npath for AM using robot manipulators that takes advantages of the robots DOF is\npresented. The path is generated based on simple surfaces in CAD models. First, the\nsurface(s) is sampled and the samples are gathered in a point cloud. Then, a path is\ngenerated based on the point cloud using a path generation algorithm. Three different\npath generation algorithms was implemented and tested: greedy choice, weighted\ngreedy choice and Travelling Salesman Problem (TSP). Out of the three algorithms,\nthe weighted greedy choice algorithm shows the most promise. With this algorithm,\npaths that enable printing along curved surfaces and reducing the need for support\nstructures was generated. The method is effective, and by interfacing with FreeCAD, it\nis easy to review the generated paths through visual aids. It is, however, important to\nmention that the method only generates paths based on simple surfaces and is based\non the assumption of fast-curing material enabling mid-air printing.