ASUS AIoT Builds New Smart Manufacturing Solutions to Promote Industrial Upgrades

Fremont, California (June 10, 2020) - With the advent of the AIoT era, ASUS has embraced new technologies and methods to develop advanced manufacturing capabilities. At the end of 2019, ASUS expanded the company's AIoT business unit to be able to develop more solutions for industries and, in the process, renamed it the AIoT Business Group (AIoT BG). After consideration and planning around the three major aspects of successful manufacturing — design power, technical ability and continuous profitability — ASUS transformed operations to achieve the flexibility, speed, productivity and quality required for supply-side Industry 4.0 upgrades.
Detecting defects by hand is a major pain point and cause of inefficiencies in manufacturing processes. By investing in smart manufacturing solutions that utilize AI for producing metal peripherals, fans, printed circuit boards and other computer components as well as for system assembly, ASUS was able to remove efficiency bottlenecks and reduce losses resulting from misjudgment of manufacturing defects by factory employees. Moving forward, ASUS will continue to use artificial intelligence and big data to statistically classify different types of quality defects, determine their causes and improve processes at the source of defects to further improve and push the boundaries of manufacturing quality.
“ASUS currently has hundreds of suppliers, and whenever we are able to improve quality-inspection processes, suppliers are receptive and willing to make changes,” said Jackie Hsu, Senior Vice President, Co-Head of Open Platform BG and AIoT Business Group said. “This is a win-win situation for ASUS and the entire industry, which has always attached great importance to product quality.”

AI Visual Inspection System

In the manufacturing industry, it is a common practice to replace manual visual inspection with automatic optical inspection (AOI). However, optical inspection is inefficient for mechanical metal parts manufacturers. Manual visual inspection often requires viewing product surfaces from multiple angles to see the defects due to the reflection of light. It is extremely important to grasp the optical and component surface characteristics to obtain complete and correct defect data.
Optical inspection is one of the core technologies of AIoT Business Group, which uses machine learning, deep learning and artificial neural network technologies to train the AI detection model correctly. "Automatic optical inspection accuracy in general is about 80–90 percent, which means that more than 10 percent of defects may be misjudged, and manual visual inspection accuracy is about 90 percent,” said Albert Chang, Corporate Vice President, Co-Head of AIoT Business Group. “At present, ASUS has enabled AI to greatly improve its accuracy to 98 percent after learning.”

AI Waveform Detection System

Fans are a key part of many computers and consumer electronics, cooling components and helping to extend product life. To ensure fan quality, manufacturers relied on inspectors who were able to detect problems with fans just by listening to them. Training highly skilled personnel for this important position took from three to six months, and inspectors would occasionally experience short- or long-term ear fatigue and other occupational factors that negatively impacted worker health and reduced the problem detection rate.
To solve this difficult problem, ASUS introduced the AI Wave Signature System, which analyzed the sounds of correctly operating fans and used them to develop a sound signature. This sound signature was then used to train AI models to qui