RS Automation, Equipped with "Brain, Muscle, and Senses," Presents a New Direction for Physical AI

On January 26, RS Automation, a company specializing in industrial motion control, announced that it is accelerating its physical artificial intelligence (AI) strategy based on "Soft Tuning," presenting a new direction in the industrial automation market.

Recently, in the field of robotics AI, models based on Vision-Language-Action (VLA) are widely utilized. VLA is an end-to-end structure that connects vision models, language models, and robot control algorithms within a single model, enabling the processing of input, understanding, and action all at once. Most VLA-based approaches require massive amounts of data and computational infrastructure, as they ensure versatility by training on large-scale unstructured data such as images and actions.

RS Automation has built a highly efficient learning structure that utilizes internal physical data from equipment, rather than collecting large-scale video and action data for training. Control signals within the equipment, such as current, position, speed, vibration, and resonance response, already reflect physical laws and directly describe the machine's state. Soft Tuning leverages signals with defined physical meaning to automatically identify resonance frequency, structural characteristics, friction, and load changes, and can estimate system characteristics even with short operation logs.

RS Automation possesses motion controllers (the "brain"), servo drives (the "muscle"), and high-precision encoders (the "senses"), and through Soft Tuning technology, synchronizes data generated from each component via EtherCAT, integrating it into dynamic data for industrial equipment. As a result, learning data naturally accumulates as the equipment is used, and the operational efficiency of the equipment itself can be automatically improved.

The company expects that if equipment performance can be improved with minimal data through the Soft Tuning model, it will lower the barriers to adoption in industrial settings and enable the rapid introduction of high-efficiency AI structures to the field.

An RS Automation representative stated, "Soft Tuning is not an AI that simply processes large amounts of data, but rather an AI that understands physics, making it a domain-specialized model focused on the field." The representative added, "Going forward, RS Automation will create a new business model distinct from conventional VLA approaches by introducing high-efficiency AI structures to industrial sites, enabling machines to autonomously improve their performance based on physical data."

원본보기 아이콘

© The Asia Business Daily(www.asiae.co.kr). All rights reserved.