Doosan Robotics has secured two national projects related to Physical AI, aiming to enhance the intelligence of collaborative robots and expand their application in manufacturing environments.
On October 6, Doosan Robotics announced that its proposals for next-generation collaborative robots and intelligent welding solutions were selected as national projects under the 'K-On-Device AI Semiconductor Technology Development Project' and the 'Robot Industry Technology Development Project,' co-hosted by the Ministry of Trade, Industry and Energy and the Korea Institute for Advancement of Technology. The total research and development budget for the two projects is approximately 989 billion won, with government support amounting to about 681 billion won.
The first project focuses on developing foundational technologies to enhance the perception, judgment, and control capabilities of collaborative robots, while the second project aims to develop applied technologies for implementing these capabilities in advanced manufacturing settings. Doosan Robotics plans to integrate these technologies to establish a 'Korean-style Physical AI Collaborative Robot Integrated Platform.'
The project on 'Next-Generation Collaborative Robot Technology Development Based on On-Device AI Semiconductors' centers on enabling robots to perceive, judge, and control independently without relying on external PCs or cloud services. This will involve integrating AI computing units (NPU) and real-time communication and control functions into domestically produced AI system semiconductors (SoC) for use in collaborative robots. The robots will make on-site operational decisions based on information from cameras, touch sensors, force and torque sensors, and voice inputs.
Doosan Robotics plans to develop three types of collaborative robots with payload capacities of 5 kg, 10 kg, and 20 kg, as well as industrial humanoids capable of dual-arm operations and omnidirectional movement. The goal is to commercialize these standard products starting in 2031 after conducting 72 hours of uninterrupted testing in at least two demand sites and obtaining international safety certification. The company also aims to expand supply to sensitive industries such as nuclear power, defense, and biotechnology.
The project duration is 54 months, with Doosan Robotics serving as the lead organization, collaborating with AIDIN Robotics, Mobilint, the Korea Electronics Technology Institute, the Korea Institute of Industrial Technology, Sage, and industry-academic cooperation groups from Korea University, Dongguk University, Sejong University, and Yonsei University.
The second project, 'Development of Intelligent Welding Solution Technology Based on Collaborative Robots and AI,' aims to automate high-level welding tasks for nuclear power equipment by combining AI and digital twin technology with collaborative robots. The robots will use cameras to recognize welding targets, correct their positions, and learn the movements of skilled workers to generate welding paths and weaving actions.
Doosan Robotics aims to align the welding paths of its robots with those of skilled workers with over 95% accuracy and reduce path deviation to within ±3 mm, as per national certification standards. The company also plans to automate multi-layer welding processes of five layers or more, reducing work time by over 50%.
Testing will be conducted at the Doosan Enerbility Nuclear Center, initially applying the technology to pre-welding, preheating, and welding processes for nuclear power equipment, followed by securing quality certification for application in shipbuilding, plant construction, and defense sectors.
The duration of this project is 45 months, with Doosan Robotics as the lead organization and DeepX, Sage, and the National Changwon University industry-academic cooperation group participating as joint research institutions. Doosan Enerbility will also participate as a demand company.
Through these projects, Doosan Robotics plans to develop Physical AI collaborative robot technology that integrates AI semiconductors, software, and robotic hardware.
Park In-won, CEO of Doosan Robotics, stated, "The labor shortage and aging skilled workforce in manufacturing are structural issues that cannot be resolved solely by expanding automation equipment. We will develop Physical AI collaborative robots that learn and replicate the skills of skilled workers by combining on-device AI semiconductors with collaborative robot platforms."
Meanwhile, Chairman Park Jung-won emphasized in his New Year's address this year that "Doosan possesses world-class manufacturing capabilities and vast hardware data in fields such as power generation equipment, construction machinery, and robotics, giving us a strong advantage in leading the era of Physical AI."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.