"Robot transformation (RX) is not just about creating a single robot; it involves designing the entire factory."
Bae Soon-min, Executive Vice President and head of the RX business team at Samsung SDS, outlined the company's vision to implement autonomy in manufacturing based on 25 years of experience in manufacturing execution systems (MES) and equipment and logistics automation. The plan focuses on connecting and integrating robots from various manufacturers into production environments.
During the 'Real Summit 2026' held on September 8, Bae introduced Samsung SDS's RX strategy under the theme of 'Robot Transformation for Autonomous Manufacturing.' The company will oversee the entire process of robot implementation in manufacturing, from process analysis and feasibility studies to robot selection, deployment, operation, and maintenance.
Samsung SDS is expanding its existing manufacturing IT capabilities into RX. The company has conducted automation projects in over 100 Samsung affiliates and more than 430 external clients over the past 25 years, leveraging its MES expertise.
Bae described the evolution of manufacturing in three stages: informatization, automation, and autonomy. He explained, "In informatization, humans make decisions and actions while machines record. In automation, humans decide and machines execute. In autonomy, AI makes decisions and robots execute."
However, he noted that physical AI and robotic technology have not yet reached completion. Bae stated, "While it is relatively easy to operate a single robot effectively, ensuring that 10 or 100 robots operate seamlessly 24/7 is challenging. What is crucial now is not just purchasing robots but building the system to support them."
Samsung SDS is developing a 'robot orchestration platform' that connects and operates robots from different manufacturers alongside existing production equipment. This platform aims to establish a stable operation system for multiple robots based on six elements: autonomy, orchestration, interoperability, resilience, safety, and security. It integrates MES with robots to streamline work instructions, data collection, and operational oversight while utilizing accumulated data for enhancing physical AI.
The company is also collaborating with external firms to build an RX ecosystem. This involves partnering with companies specializing in robot hardware, physical AI, data, simulation, system integration, and maintenance to combine necessary technologies for manufacturing environments. Currently, 10 companies, including Rainbow Robotics and Walden Robotics, are participating in this initiative.
Bae proposed the 'DRT' methodology for applying RX in manufacturing settings. This four-step approach includes 'Discover,' which involves analyzing processes and tasks on-site; 'Assess,' which evaluates the feasibility of robot application based on current technology; 'Redefine,' which may involve redesigning processes or robots if necessary; and 'Transform,' which encompasses deployment, evaluation, stabilization, and maintenance.
He emphasized, "The journey of RX does not start with purchasing robots; it begins with analyzing the site and identifying customer issues to determine which processes are suitable for robot application."
Following the session, various robots, including Rainbow Robotics' quadrupedal and humanoid robots, were showcased. Ahn Dae-jung, Vice President in charge of digital factory at Samsung SDS, explained, "Rainbow Robotics' quadrupedal robot and mobile humanoid robot RB-Y1 are being deployed in project units within Samsung affiliates."
According to Ahn, Walden Robotics' robots are also being utilized in actual production lines, operating around the clock, while Samsung SDS is validating the on-site applicability of related robots based on use cases from Samsung affiliates.
Samsung SDS plans to apply various types of robots based on the manufacturing environment and task characteristics, rather than focusing solely on humanoids. Ahn stated, "Humanoid physical AI is not necessarily the best option; it is essential to analyze the robot's form factor, learning methods, and operational costs comprehensively before making a selection."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.