LG Electronics: "Spaces Are Robots... Appliances Must Share 'Sensor Data'"

SungHyuk Kim Delivers Keynote at Advanced Sensor Forum
Standalone Sensor Performance Is No Longer Enough
Integration of Intelligent Sensors and LLMs Is Essential
"Korea Can Lead Globally by Linking Home Appliances, Robots, and Manufacturing Ecosystems"

In order to succeed in the domain of physical artificial intelligence (AI), it is necessary to move beyond the development of humanoid (human-shaped) robots and instead consider entire spaces—such as homes and factories—as single robots, building an integrated system in which appliances within those spaces organically share sensor data.

Seonghyuk Kim, Executive Director at LG Electronics, is delivering the keynote speech titled "Physical AI Home Robot Era, Sensor Strategy for Space Recognition" at the 11th Advanced Sensor Forum held on the 15th at the Chosun Palace Hotel in Gangnam-gu, Seoul. Photo by Hyunji Kwon

SungHyuk Kim, Executive Director of LG Electronics, emphasized this point during his keynote speech at the 11th Advanced Sensor Forum, held on the 15th at the Chosun Palace Hotel in Gangnam-gu, Seoul, under the theme "Physical AI Home Robot Era, Sensor Strategy for Space Recognition."

Executive Director Kim noted that, in the era of physical AI, the traditional concept of a robot should be expanded to encompass entire spaces. He explained that while a robot is generally made up of three core components—sensors, computation/brain, and actuators—expanding this notion means that any space where people live or work, such as a home or factory, can itself become a robot. He said, "When people think of physical AI, most tend to focus only on humanoid robots. However, a home can also perform the functions of a robot if it is equipped with sensors, a brain, and actuators," adding, "Rather than simply equipping individual robots with high-resolution vision sensors to boost performance, sensor systems should be designed from the perspective of recognizing the entire space in order to realize the commercialization of physical AI."

He also suggested that, rather than competing over the performance of standalone sensors, the market can be led by building a 'closed loop' system in which devices share and verify the data sensed and the actual state of the environment. "For example," Executive Director Kim explained, "even if a robot vacuum cleaner does not have its own sensors, it can avoid obstacles and locate cleaning areas by using positional and sensor data shared from other nearby appliances such as washing machines or air conditioners. Instead of focusing all capacity on a single, high-performance sensor, it is crucial that the roles of each sensor are divided so that sensations such as touch, temperature, and material can each be individually learned."

He further analyzed that the paradigm in sensor technology is also rapidly shifting from the 'always-on' stage, which focused on simple on-off control or power adjustment, to the 'always-aware' stage that aims to understand the context and situation of the field. This means sensors should go beyond simply detecting “some gas is leaking” or “there is an odor” and instead provide contextual information that humans can immediately interpret and act upon, such as identifying whether it is “the smell of garbage” or “a pet’s excretion.”

Executive Director Kim highlighted that it will become increasingly important to group this intelligent sensor data over time, connect it to large language models (LLM), and create AI-based models for sensors that seamlessly integrate on-device AI and the cloud.

He emphasized, "Korea has all the components of the physical AI ecosystem—home appliances, robots, manufacturing capabilities, and sensor parts. As long as we integrate these well, we can take the lead."

This content was produced with the assistance of AI translation services.