Fixing 'Strange Noises' in Factory... De...

[Startup Story] Jaehyun Lee, CEO of LOAS

Author Malcolm Gladwell introduced the concept of “Thin-slicing” in 2005. It refers to making instant judgments based on only small clues. Representative examples include an experienced tennis coach predicting a double fault the moment they observe a player’s serve motion, or a firefighter instinctively sensing that a building is about to collapse during a fire and immediately ordering evacuation for reasons difficult to explain. The idea is that these are not simple intuitions, but conclusions reached instantaneously through years of accumulated experience and training. It is also referred to as “trained intuition.”

There is a group of experts in industrial environments who possess this kind of intuition. In manufacturing sites, each factory typically has five to ten specialists capable of identifying defects or abnormalities simply by listening to the sounds of operating equipment.

Jaehyun Lee, CEO of LOAS, traveled extensively across numerous domestic and international industrial sites while working as an engineer. During that time, he began questioning, “Even in the 21st century, should humans still be doing this manually?”

That question ultimately led him to develop “AI Square,” an algorithm capable of listening to sounds and predicting risks in advance in place of human operators.

Conventional acoustic inspections in industrial environments, which relied heavily on experienced personnel, faced inherent limitations in eliminating subjective judgment. While some industrial sites attempted to introduce specialized equipment, inspections typically required production facilities to be halted during the process. Simply stopping factory operations could result in financial losses amounting to hundreds of millions of won.

Jaehyun Lee stated, “Even in an era where automation has advanced this far, the sight of people standing on production lines carefully listening for noise is, in itself, quite absurd.” He added, “There has always been market demand for quantifying this type of acoustic diagnosis, but most attempts failed because they could not distinguish between the surrounding noise inside factories and the actual fault sounds that mattered.”

The core of LOAS’s proprietary algorithm, “AI Square,” lies in its ability to isolate only the sounds it needs. It works much like how parents can instantly recognize their child calling “Mom” within a noisy crowd.

The system first divides spaces into multiple zones using numerous multi-channel microphone sensors. After collecting all sounds within those spaces, any sound identified as noise is intentionally suppressed during the collection process. It functions as a software-based acoustic filtering system.

At the same time, cameras visually identify target objects and focus on collecting surrounding sounds. For example, if the inspection target is factory piping, the system amplifies sounds specifically along the contour area corresponding to those pipelines, allowing the AI to perform detailed inspections.

This algorithm, “AI Square,” is embedded into multiple LOAS solutions, including the intelligent acoustic inspection system “SMART,” the industrial integrated monitoring platform “ARQOS,” and the urban drone detection system “AirScope,” all of which are responsible for industrial safety monitoring.

All three products detect and analyze abnormal acoustic data within noisy industrial environments and provide real-time location information for detected abnormalities.

Technology Validated by LG Electronics and Samsung Electronics — “Superior to Human Capability”

The technology developed by LOAS has already proven its capabilities within the production environments of major corporations including Samsung Electronics and LG Electronics. Currently, LOAS systems are deployed on LG Electronics home appliance production lines, replacing human inspectors to detect equipment defects.

The system continuously improves its accuracy by retraining on more than 60,000 data cases every day. Thanks to this accumulating data, LOAS’s algorithm has evolved beyond real-time diagnostics to a level capable of predicting when and how equipment failures may occur in the future.

Jaehyun Lee stated, “Technology research divisions at companies such as LG directly compared the detection performance of human inspectors and AI Square, and concluded that AI Square demonstrated superior performance.” He added, “By immediately deploying LOAS products into industrial sites, companies have been able to reduce diagnostic personnel by approximately 5 to 10 workers per production facility.”

The drone system “AirScope” is already operating at Samsung Electronics facilities in Giheung and Hwaseong, and will expand to the Pyeongtaek campus this coming June.

Korea Western Power is also utilizing LOAS products. Korea Western Power collaborated closely with LOAS by providing critical field data that became essential to the company’s technological advancement.

Lee stated, “We were selected for a public data utilization project, which allowed us to receive data directly from Korea Western Power plant sites and further advance our technology.” He added, “Because power plants are national critical infrastructure, the procedures were extremely strict, even involving oversight from the National Intelligence Service. But without Korea Western Power, it would have been very difficult for LOAS to validate its technology.”

Targeting the Middle East and West Asian Markets — “Our Goal Is a Human-Free Factory”

The customer base of LOAS is rapidly expanding. The company has begun supplying mass-production products to cathode material factories operated by EcoPro. LOAS has also launched new projects with Korea East-West Power, while automotive seat manufacturers such as Hyundai Transys, an affiliate of Hyundai Motor Company, are increasingly adopting LOAS solutions.

Driven by this momentum, LOAS has set its minimum revenue target for this year at over KRW 5.5 billion.

Beyond manufacturing facilities and power plants, the company is also preparing to enter the urban security market in the future. One example is a “patrol robot” capable of autonomously navigating toward the exact direction and location of a person screaming in public spaces. This represents the expansion of technologies refined within industrial environments into broader urban safety applications.

LOAS is also targeting global expansion into the Middle East and West Asian markets, where the plant industry is particularly large. The company is considering leveraging overseas plant projects conducted by state-owned energy corporations as a pathway into those markets.

At the same time, LOAS plans to further advance its acoustic diagnostic technologies through joint research projects with Fraunhofer-Gesellschaft in Germany.

Jaehyun Lee stated, “By conducting safety diagnostics without deploying personnel into dangerous environments, companies can improve production quality while also strengthening manufacturing cost competitiveness.” He added, “Our goal is to transform customer factories into true ‘dark plants’ capable of operating entirely without lights or human personnel.”

Source: Money Today > Reporter Choi Woo-young

View Full Article>>https://n.news.naver.com/article/008/0005358946

─

LOAS Inc.

(주)로아스

Contact

Tel. 02-6486-6411

Email. info@loas.ai