Unichem subsidiary Lumia accelerates development of ‘tactile sensors’ to teach robots human dexterity

[Financial News]  Lumia Technologies, a U.S. flexible-sensor company for which Unichem has signed an agreement to acquire a 100% stake, is accelerating development of technology that turns human hand skills into data for robot training. The plan is to collect data on the pressure workers apply and how they control force in logistics settings, and use it to improve robots’ precision manipulation capabilities. Attention is focused on whether the company can expand its revenue model through a tactile-data business in the future.
According to industry sources on the 22nd, Lumia recently completed the design of the key sensor for a smart glove that analyzes logistics-center workers’ hand movements and began performance testing.
The core of the technology is to capture data on the subtle pressure and force adjustments people make when picking up or moving objects. The data can be used to train robots to adjust the force of their grip according to an object’s weight and shape, the company explained.
Lumia has combined its proprietary flexible electronic circuit (LEL) and pressure sensor (FSR) technologies. It is developing a fabric-based sensor structure designed for prolonged repetitive tasks, taking into account eight-hour shifts at industrial sites and loads in the 50–200 N range.
The ongoing assessment is at an initial validation stage, checking the sensor’s measurement stability, wash durability, and detection performance in curved areas. Based on the results, the company plans to improve the glove structure and assess its potential for use in the field.
Lumia has also previously developed a tactile-sensor glove prototype for use with the robotic hand of Festo, a German industrial automation company. The prototype incorporated 131 sensing points.
In the robotics industry, analysts say the importance of acquiring tactile data could grow as the physical AI race expands from visual recognition to the precise manipulation of real-world objects. One industry source said, “Pressure and force-control data collected in industrial settings can be used to improve the accuracy of robot training, making the commercialization of related technologies a factor that could determine competitiveness.” 
Meanwhile, alongside commercializing its smart glove, Lumia plans to commercialize the tactile data collected through it. It has identified humanoid manufacturers, robotic-arm developers, and AI-related companies in South Korea and the United States as potential customers.
Lumia said, “For humanoid robots to perform precise tasks in industrial settings, high-quality data reflecting human touch and force control is needed,” adding, “Through sensor validation, we will establish a foundation for commercializing technology to collect data for robot training.”


[email protected] Kim Gyeong-a Reporter