Fleet operated with US firm Applied Intuition

Operations centered on Seoul, Incheon and Gyeonggi Province

Fleet to expand to 20 vehicles by 2028

Vehicles equipped with company's own cameras, radar and lidar

Real-world driving data to sharpen sensor fusion technology

Company accelerates push into physical AI market

An LG Innotek data collection vehicle for autonomous driving sensing travels through an urban area. [LG Innotek]
An LG Innotek data collection vehicle for autonomous driving sensing travels through an urban area. [LG Innotek]

LG Innotek has begun operating vehicles on domestic roads to collect sensing data for autonomous driving. Equipped with cameras, radar and lidar, the vehicles gather real-world driving data to advance the company's sensor fusion technology and accelerate its push into the physical AI market, including autonomous driving.

LG Innotek said Tuesday it is operating autonomous driving sensing data collection vehicles in South Korea together with Applied Intuition, a US autonomous driving software company. The two companies signed a strategic partnership in March to advance autonomous driving sensing solutions.

The data collection vehicles will operate primarily in Seoul, Incheon and Gyeonggi Province. LG Innotek plans to add vehicles dedicated to autonomous driving software development and expand its domestic fleet to about 20 by 2028.

It is rare in the industry for an autonomous driving sensing solution provider to operate its own data collection fleet. LG Innotek said it will use data gathered on actual roads to improve the performance and reliability of individual sensors — including automotive cameras, radar and lidar.

Each data collection vehicle carries multiple sensors, including cameras, radar and lidar, and continuously collects real-time information on road conditions, vehicle movement, object positions and weather changes while driving. The data is then combined with digital twin virtual environments to improve sensor performance and validate systems.

LG Innotek is particularly focused on using real-world driving data to advance its proprietary sensor fusion technology. Sensor fusion combines image data from cameras, distance and speed information from radar, and three-dimensional spatial data from lidar to enable precise perception of the surrounding environment. As autonomous driving grows more sophisticated, the ability to synthesize data from multiple sensor types — not just improve individual sensors — is becoming increasingly important.

The fleet will also capture edge case data — unusual scenarios that can pose problems for fully autonomous driving. These include black vehicles in heavy rain, strong backlighting at tunnel exits and irregular lane markings at construction sites, all of which are difficult to encounter in ordinary conditions. Personally identifiable information captured during data collection, such as pedestrian faces and vehicle license plates, will be anonymized.

LG Innotek plans to use the fleet data to strengthen its competitiveness in integrated sensing solutions spanning automotive cameras, radar and lidar. "There are very few companies in the world that can supply all three types of modules with their own technology," the company said.

Applied Intuition data collection vehicles equipped with LG Innotek sensing modules are also set to operate in the United States, Europe and Japan. By analyzing large volumes of real-world autonomous driving data, LG Innotek aims to identify areas for improvement across diverse environments in advance and proactively incorporate them to maximize module performance.

Min Joon, LG Innotek's chief technology officer, said autonomous driving solutions are evolving beyond improving individual sensor performance toward jointly optimizing sensors and software based on data. "We will quickly establish a leading position in the physical AI sensing market, including autonomous driving, by leveraging advanced integrated sensing solutions built on data," he said.

LG Innotek is cultivating autonomous driving and robotics as new growth engines under its physical AI strategy. Last year the company signed a strategic partnership with US lidar firm Aeva, and this year it joined forces with Applied Intuition and Kakao Mobility, South Korea's largest mobility platform, to co-develop autonomous driving solutions. It also supplies camera modules for Zoox, Amazon's autonomous vehicle subsidiary, for use in its robotaxi.


go@heraldcorp.com