"Immediately Detecting Disaster Situations by Sound Alone": Korean AI Technology Attracting Global Attention

[Startup Drinks & Talks] Interview with LOAS, an AI-Based Acoustic Inspection System Developer

A rusty bicycle makes a creaking sound. When unfamiliar noises come from a car engine, it signals the need for inspection. Abnormal sounds are, in essence, a “signal” that something is wrong.

The same applies in industrial environments. Situations such as electrical leakage or gas leaks generate sounds outside the normal range. The challenge is that surrounding environmental noise and the limitations of human hearing make it difficult to accurately capture every sound. Problems that go undetected in time can escalate into major risks, including safety accidents and an increase in defective products.

Jaehyun Lee, CEO of startup LOAS, believed that sound could solve a wide range of problems occurring in industrial sites. This became the motivation behind developing AI-based acoustic detection, tracking, and inspection solutions. By deeply understanding on-site challenges, LOAS rapidly expanded its solutions into major corporations and public enterprises. We met with Lee to discuss the role sound plays in industrial environments.

◇Clues in Industrial Sites Lie in Sound

Lee majored in physics at university and continued on to a master’s degree in optics. After graduating from graduate school in 2016, he joined a company manufacturing optical in-vitro diagnostic systems. Through the foreign company, he worked in Germany as a field application engineer applying measurement and inspection systems to industrial plants.

There, he realized there was significant demand for diagnosing noise generated during industrial processes. “Noise and vibration generated in industrial environments are directly connected to productivity and working conditions. Abnormal sounds and vibrations serve as signals of equipment failure or danger detection. Even sounds too faint for humans to hear can still create discomfort simply by existing. The key issue is detectability. For vibration, commercialized vibration meters and sensors already exist and are widely used. However, noise diagnosis technology remains relatively underdeveloped.”

To eliminate noise, the source of the noise must first be identified. “To determine whether a product has abnormalities, noise diagnosis is typically performed in an anechoic chamber where environmental noise is blocked out. But this approach has many limitations when applied to actual production environments. As a result, there are still many cases where abnormal noise is judged directly by human hearing. Devices such as microphone sensors are also used, but industrial environments are extremely noisy, making it difficult to distinguish whether a sound originated from equipment or products, or from surrounding environmental noise.”

Lee believed that a system capable of accurately detecting the source and type of noise directly within industrial sites would have enormous impact. “Acoustics are the most intuitive physical means of showing where a problem occurs. Vibration detection sensors can indicate that an abnormality exists, but they cannot identify exactly where it is happening. However, when gas leaks occur in pipelines or electrical leakage happens, acoustic data becomes the decisive clue for identifying the cause. If we extend beyond the audible range into ultrasonic regions humans cannot hear, we can detect surrounding problems with even greater precision.”

<Excerpt Omitted>

The journey to this point was far from easy. “For the first two and a half years focused entirely on technology development, we struggled financially. Last year was also a critical turning point. Before participating in an exhibition, I even thought, ‘This could be the end of our company.’ Every team member worked day and night preparing for the exhibition. Thanks to everyone’s united effort, we were able to secure investment and advance our products, creating a foundation for growth. Looking back, we were fortunate. But I also believe our ability to make quick decisions and immediately execute them played an important role. We turned things people in industrial sites could only imagine into reality.”

The true stage for LOAS lies overseas. The company is currently developing robots tailored to individual power plants in order to introduce maintenance solutions to overseas plants operated by Korea Western Power. Two major corporate clients are also planning to deploy the solutions across their overseas facilities.

“We are building domestic references as a bridgehead for entering global markets. Fortunately, we have secured strong references, and based on them, we are knocking on the doors of overseas markets. Our targets are China, the Middle East, and Southeast Asia, where manufacturing industries are highly developed. We aim to establish ourselves as a leading company in the field of acoustic diagnostics.”

Source: Chosun Biz > Startup Drinks & Talks / Reporters Park Yu-yeon and Jin Eun-hye, THE B·BID

View Full Article>>https://www.chosun.com/economy/startup_story/2025/04/22/SDK3L3J4OFCSBC3DOAJF75CNE4/

─

LOAS Inc.

(주)로아스

Contact

Tel. 02-6486-6411

Email. info@loas.ai