Machine learning-driven 3D printing: A review

In recent years, Addit. Manuf. (AM, i.e. 3D printing) has been developing rapidly as a much-needed strategic technology in the field of high-end intelligent manufacturing. However, it is constrained by the cross-influence of materials, processes, and algorithms, which make the print quality more varied, and fortunately machine learning has brought new optimization strategies for 3D printing. This paper provides an overview of machine learning-driven 3D printing technology, with a particular focus on its process, monitoring, and motion planning. In addition, the introduction of a 6-DOF robotic arm allows for more versatile printing paths. We found that supervised learning has a high rate of application in process optimization and surface quality inspection through the review. While reinforcement learning and deep learning have a clear advantage in path planning for multidegree of freedom printing platforms, which sets the stage for pioneering applications of AI in Addit. Manuf..