"Robot Controlled by Brain Signals" World's First 'Brain-to-Robot' Development Begins

Development of a next-generation brain-robot interface platform that enables interaction between the human brain and exoskeleton robots has begun in South Korea. The core of this platform is to enable real-time control of a robot's movements via signals sent from the brain, while simultaneously allowing tactile and force feedback detected by the robot to be transmitted back to the brain.

KAIST announced on June 25 that the research team led by Professors Kong Kyungchul and Kim Joung from the Department of Mechanical Engineering, together with Angel Robotics Co., Ltd., have launched the world's first bidirectional "Brain-to-Robot" system. This project, a flagship initiative under the pan-ministerial Advanced Medical Device R&D Program, began in April and will continue until December 2032.

Hierarchical Robot Control Architecture Mimicking Human Motor Mechanisms. KAIST

Brain interface technology, which uses brain signals to move a cursor or control a smartphone, has already reached the stage of clinical trials in humans. Global companies such as Neuralink (USA) and Synchron are currently accelerating the development of related technologies.

However, existing technology faces limitations in connecting actual movement and sensation simultaneously. Because it is unclear what is being controlled by the brain signals and what sensory feedback is being returned, research so far has focused on advancing signal decoding technology.

In contrast, "Brain-to-Robot" directly targets the exoskeleton robot for control. The aim is to create a bidirectional interface in which the user's brain signals, reflecting their intent to act, are used to move the robot, while feedback such as ground reaction force (the force the ground exerts against the foot), joint torque (rotational force at the joint), and tactile information sensed by the robot are transmitted back to the brain.

No bidirectional "Brain-to-Robot" system has been reported anywhere in the world. This is why the current research is expected to be a new turning point in brain interface technology.

During the development of "Brain-to-Robot," the KAIST research team is responsible for developing the core technologies. Professor Kong's team is tasked with developing wearable robot control and AI-based motion intention interpretation technology. A core focus is the design of a somatosensory interface (a system for transmitting bodily sensations) to deliver sensory information detected by the robot to the Brain Chip (a semiconductor for processing brain signals). Professor Kim's team is working on developing robotic skin that can sense on behalf of people with disabilities and AI-based somatosensory interpretation technology.

The research teams are also developing AI-based encoding and decoding algorithms to convert brain signals into robot control commands and transmit sensory information detected by the robot back to the brain.

The main challenge lies in processing hundreds of channels of cortical signals (neural signals generated in the cerebral cortex) in real-time and reliably maintaining a closed-loop system (a structure in which signals are exchanged in real-time with extremely short latency).

AI-generated image. KAIST

Commercialization will be handled by Angel Robotics (KOSDAQ:455900), a company founded by Professor Kong. Ultimately, the joint research team plans to pursue full-cycle commercialization of "Brain-to-Robot," from approval by the Ministry of Food and Drug Safety to actual deployment.

Professor Kong stated, "'Brain-to-Robot' will offer a new rehabilitation paradigm that enables people with quadriplegia to walk independently and grasp objects in daily life outside of the hospital, even feeling sensations at their fingertips."

The research team emphasized that, as "Brain-to-Robot" is an unprecedented high-difficulty convergence technology both in Korea and globally, ensuring long-term safety, clinical validation, and regulatory frameworks must proceed in parallel with technology development. They also highlighted that, for entry into the global market, the system design must organically integrate safety and efficacy verification, accumulation of clinical evidence, establishment of risk management systems, brain signal data protection and cybersecurity, as well as review for ethical acceptability.

Lee Kwanghyung, President of KAIST, said, "The 'Brain-to-Robot' flagship project is one of the most challenging convergence research initiatives in the world, led by KAIST's research team. We will fully leverage KAIST's strengths in brain interface, AI, semiconductors, and robotics research to drive next-generation innovation in 'Brain-to-Robot' technology."

Professor Kong, who founded the gait-assist exoskeleton robot company Angel Robotics, is a world-renowned researcher in the field of wearable robotics, having won gold medals twice consecutively at the international Cybathlon (an international competition for assistive technology for people with disabilities). Professor Kim is also recognized as a leading researcher, having received the "Scientist of the Month" award for his research on robotic skin technology.

The research teams led by these professors have now formed a consortium and are actively pursuing research on the "Brain-to-Robot" platform, which combines brain neural interfaces with exoskeleton robots.

This content was produced with the assistance of AI translation services.