Time-optimal trajectory generation, path planning and control for a wood patching robot

Patching of wood defects of shuttering panels, e.g. loose dead knots, is a costly manual process of inline production in timber industry. The need for automation is apparent. Given a list of patch positions for a panel, time-optimal trajectories between the individual positions are generated. Then, their time-optimal processing sequence is determined, a problem similar to the traveling salesman problem. It is solved using the so-called Local Search Receding Horizon Algorithm. The focus of this paper is set on the real-time position control of the patching robot. Since the panel is moved by the patching robot via frictional grip, it is likely that slip occurs between the panel and the patching robot. This unknown relative motion has to be corrected. However, an absolute position measurement of the panel is only available at a very low sampling rate. Therefore, a control strategy with a reference trajectory update is proposed which allows for precise and fast positioning of shuttering panels.