Optimal Path Planning for Painting Robots

In this work, a robotic painting task is addressed in order to automate and improve the efficiency of the process. Usually, path planning in robotic painting is done through self learning programming. Recently, different automated and semi-automated systems have been developed in order to avoid this procedure by using a CAD-drawing to create a CAD-guided trajectory for the paint gun, or by acquiring and recognizing the overall shape of the object to be painted within a library of prestored shapes with associated pre-defined paths. However, a general solution is still lacking, which enables one to overcome the need for a CAD-drawing and to deal with any kind of shapes. In this paper, graph theory and operative research techniques are applied to provide a general and optimal solution of the path planning problem for painting robots. The object to be painted is partitioned into primitives that can be represented by a graph. The Chinese Postman algorithm is then run on the graph in order to obtain a minimum length path covering all the arcs (Eulerian path). However, this path is not always optimal with respect to the constraints imposed by the painting process, hence dedicated algorithms have been developed in order to generate the optimal path in such cases. Based on the optimal path, the robot trajectories are planned by imposing a constant velocity motion of the spray gun, in order to ensure a uniform distribution of the paint over the object surface. The proposed system for optimal path planning has been implemented in a Matlab environment and extensively tested with excellent results in terms of time, costs and usability.