The many tunable parameters involved in laser processing, such as wavelength, pulse duration, pulse energy, and scan speed, not to mention various other complicating factors on the material side, makes it practically impossible to reliably find an optimized parameter set to realize a specific processing target. Currently, an acceptable parameter set is mainly found by tapping the experience and intuition of skilled people within the present production system. However, such methods do not scale to the mass-customization needs of the coming super-smart society, and it has become critical to develop ways to transfer such human experience and intuition to a more scalable setting: namely, the cyber-space. A major challenge in developing a cyber-space solution has been augmenting the limited experimental and theoretical insights of the laser processing phenomenon to the specific problems at hand. Here, we focus on automated data acquisition systems coupled with artificial intelligence (AI) methods to overcome this technological gap. We propose ways to realize cyber-physical systems specializing in specific facets of laser production by showing experimental results from four kinds of automated data acquisition systems. We lastly discuss such methods in context as an important first step to creating an AI based cyber-physical simulator.

中文翻译：

---

激光生产信息物理系统的全自动数据采集


激光加工涉及许多可调参数，例如波长、脉冲持续时间、脉冲能量和扫描速度，更不用说材料方面的各种其他复杂因素，使得实际上不可能可靠地找到优化的参数集来实现特定的目标。处理目标。目前，可接受的参数集主要是通过利用现有生产系统中技术人员的经验和直觉来找到的。然而，此类方法无法满足即将到来的超级智能社会的大规模定制需求，因此开发将此类人类经验和直觉转移到更具可扩展性的环境（即网络空间）的方法已变得至关重要。开发网络空间解决方案的一个主要挑战是将激光加工现象的有限实验和理论见解扩展到当前的具体问题。在这里，我们专注于自动化数据采集系统与人工智能（AI）方法相结合，以克服这一技术差距。我们通过展示四种自动数据采集系统的实验结果，提出了实现专门从事激光生产特定方面的网络物理系统的方法。最后，我们在上下文中讨论这些方法，作为创建基于人工智能的网络物理模拟器的重要第一步。