If you ever visit a smart factory, you can see IoT sensors and cameras virtually everywhere. These gadgets collect, store and analyze data. The analyzed data serves to pinpoint product design flaws and identify any facilities exhibiting symptoms of malfunction. This, in turn, helps provide an accurate snapshot of an entire smart factory where each and every facility and device is connected through wireless communication to allow for the real-time analysis of the whole process. This emerging technology is garnering attention from such renowned manufacturers as Samsung Electronics, Hyundai Motor Company, SK Hynix, and Hyundai Heavy Industries.

The Factory Intelligence Laboratory (also known as FiLab), directed by professor Duck Young Kim at the Department of Industrial & Management Engineering, POSTECH, delves into the core technologies required to enable smart factories. The scope of research is truly extensive, from operating system platforms aligned with international standards for smart factory control, operation and management to intelligent assembly line robots, machine learning-based facility maintenance and even Digital Twin technology that creates a virtual replica of its physical entity.

Researchers at the Lab are zeroing in on smart factory technologies that can enable the manufacturing of many kinds of products in extremely small batches at price points on par with those that are mass produced.  Conventional factories, by their default design, are able to mass produce a single product and are equipped with facilities to churn out a specific type of product at an affordable price. While this typical manufacturing approach is efficient from both the time and cost perspective, it often leaves much to be desired in producing diverse low-volume products.

At FiLab, work is underway to overcome such limitations through rapid factory reconfiguration and intelligent assembly robotics: the former quickens the pace of facility configuration to switch to the manufacture of different products while the latter helps modify the product assembly method accordingly. The Lab has already established its own smart factory testbed, the largest-ever among peer universities, to develop and verify associated core technologies. FiLab is also working on ways to broaden the scope of communications within the smart factory and methods to compensate for robotic errors.

To explore ways to maintain facility functionality and quality, researchers are probing sensor data interpretation technology to enable predictive maintenance and quality inspection through machine learning. The data they collect and analyze is actual data gathered on the shop floor, and therefore very much unrefined. This has allowed the Lab to develop systems to support the diagnosis of abnormal operations on ships and automotive engines and analyze the root causes of dysfunction, real-time control and compensation in the laser welding process, and noise detection and defect location. These research outcomes were featured in an international academic journal published by the Institute of Electrical and Electronics Engineers (IEEE).

FiLab aims to develop and unveil a smart factory operating system conformant with applicable international standards and designed for the purpose of the efficient control, management and operation of smart factories. Akin to Lixus, an open source computer operating system platform, the FiLab is working on an easily accessible smart factory operation system that can seamlessly be utilized all around the world.