Previous algorithms for slicing, path planning or trajectory planning of additive manufacturing cannot be used consistently for multidirectional additive manufacturing with pure object manipulation in wire-arc additive manufacturing. This work presents a novel path planning approach that directly takes robot kinematics into account and thus ensures the reachability of all critical path poses. In an additional step, the planned path segments are smoothed so that joint velocity limits are respected. It is shown that the implemented path planner generates executable robot paths and at the same time maintains the process quality (in this case, sufficient coverage of the slice area). While the introduced method enables the generation of reachable printing paths, the smoothing algorithm allows for the execution of the path with respect to the robot’s velocity limits and at the same time improves the slice coverage. Future experiments will show the realization of the real robot setup presented.

Wire Arc Additive Manufacturing (WAAM) with eccentric wire feed requires defined operating conditions due to the possibility of varying shapes of the deposited and solidified material depending on the welding torch orientation. In consequence, the produced component can contain significant errors because single bead geometrical errors are cumulatively added to the next layer during a building process. In order to minimise such inaccuracies caused by torch manipulation, this article illustrates the concept and testing of object-manipulated WAAM by incorporating robotic and welding technologies. As the first step towards this target, robotic hardware and software interfaces were developed to control the robot. Alongside, a fixture for holding the substrate plate was designed and fabricated. After establishing the robotic setup, in order to complete the whole WAAM process setup, a Gas Metal Arc Welding (GMAW) process was built and integrated into the system. Later, an experimental plan was prepared to perform single and multilayer welding experiments as well as for different trajectories. According to this plan, several welding experiments were performed to decide the parametric working range for the further WAAM experiments. In the end, the results of the first multilayer depositions over intricate trajectories are shown. Further performance and quality optimization strategies are also discussed at the end of this article.

The recognition of human activities from video sequences and their transformation into a machine-readable form is a challenging task, which is the subject of many studies. The goal of this project is to develop an automated method for analyzing, identifying and processing motion capture data into a planning language. This is performed in a cooking scenario by recording the pose of the acting hand. First, predefined side actions are detected in the dataset using classification. The remaining frames are then clustered into main actions. Using this information, the known initial positions and virtual object tracking, a machine-readable planning domain definition language (PDDL) is generated.

Sparse and imbalanced data is a common challenge that practitioners must overcome when implementing industrial ML applications. This challenge concerns deep learning-based quality inspection systems in particular, as they often are obligated to adhere to high constraints in terms of reliability and performance. As deep learning quality inspection systems are usually implemented in a supervised manner, they additionally require balanced datasets that may be difficult or costly to obtain in production environments. However, new approaches using Generative Adversarial Networks for synthetic image generation promise a remedy by increasing the data amount of sparse classes, such as faults or defects. This paper presents an experimental use case where we employ a state-of-the-art image generator model of StyleGAN2 to a quality inspection application in laser beam welding to increase the number of defect images for training an object detector. We evaluate the generated images and their influence on the object detector’s performance using several training configurations. Our results reveal that with the limited amount of data, we are able to generate synthetic images that look promising at first glance. However, in the evaluation based on the object detector, we find that introducing synthetic images had an adverse effect on detection performance and robustness of the system. Further research is required to generate defect images from sparse datasets that can improve the performance of object detection systems in quality inspection.