Material Extrusion is the fundamental idea behind Fused Deposition Modeling, the most widespread Additive Manufacturing (AM) method

材料挤出是熔融沉积成型背后的基本思想，熔融沉积成型是最广泛的增材制造（AM）方法。它可以在内部快速成型以前不可行的几何形状，加速生物医学，航空航天等行业。缺点是，该技术是有限的，因为它容易出错，主要是以低尺寸和几何公差的形式。更严重的缺陷也很常见，如翘曲，过度挤压，开裂和拉丝。通常情况下，一个人类工作者是必不可少的，以监测过程，因为它发生，并遵循一个试错的方法，每当出现问题。这一要求大大增加了运营成本，而且耗时，阻碍了技术的更广泛适应。建议的项目将旨在研究人工智能如何帮助自动化所涉及的重复性任务，同时受益于AM的数字性质。当前的研究重点是最近的一项突破，即CAXTON，该突破可以使用安装在多个联网3D打印机上的图像传感器对挤出过程进行现场监控。这使得快速和多样化的数据收集成为可能。通过在收集的数据上应用成熟和最先进的机器学习模型，将探索不同的方法，并报告其性能。

Material Extrusion is the fundamental idea behind Fused Deposition Modeling, the most widespread Additive Manufacturing (AM) method. It enables in-house rapid prototyping of previously infeasible geometries, accelerating industries such as biomedical, aerospace. On the downside, the technology is limited as it is prone to errors, mostly in the form of low dimensional and geometric tolerances. More sever defects are also common, such as warping, over-extrusion, cracking and stringing. Usually, a human worker is essential, to monitor the process as it happens and follow a trial-and-error approach whenever something goes wrong. This requirement significantly adds to the operating costs and is time consuming, hindering the wider adaptation of the technology. The suggested project will aim to investigate how Artificial Intelligence can help automate the involved repetitive tasks while benefiting from AM's digital nature. Current research focuses on a recent breakthrough known as CAXTON, which enables the in-situ monitoring of the extrusion process using image sensors installed on multiple networked 3D printers. This enables quick and diverse data collection. By applying well established and state-of-art machine learning models on the collected data, different approaches will be explored and their performances will be reported.