Key Takeaways
- KAIST professors Kyoungchul Kong and Jung Kim are leading a world‑first bidirectional “Brain‑to‑Robot” platform that links an exoskeleton to human brain signals for real‑time control and sensory feedback.
- The project runs from April 2026 to December 2032 under the Korea Medical Device Development Fund (KMDF) and involves Angel Robotics, the company founded by Prof. Kong.
- Core technologies include wearable‑robot control, AI‑based movement‑intention decoding, a somatosensory interface, and robotic skin that returns ground‑reaction force, joint torque, and tactile data to the brain.
- A major technical hurdle is processing hundreds of cortical signal channels while maintaining an ultra‑low‑latency closed loop for seamless brain‑robot communication.
- Commercialization will pursue full regulatory approval, clinical validation, and real‑world deployment, aiming to enable people with quadriplegia to walk, grasp objects, and feel sensations in daily life.
- Safety, efficacy, data privacy, cybersecurity, and ethical review must advance in parallel with the technology to reach global markets.
- KAIST’s broader brain‑interface research spans neural‑intention recognition, ultra‑low‑power bio‑neural circuits, high‑resolution neural electrodes, AI‑semiconductor systems, and precision brain‑signal measurement.
Project Overview and Objectives
KAIST has launched a flagship initiative to create the world’s first bidirectional Brain‑to‑Robot system. By decoding movement intentions from cortical brain signals, the platform drives an exoskeleton in real time, while simultaneously transmitting the robot’s tactile and force sensations back to the user’s brain. The goal is to establish a complete closed‑loop interface that restores not only motor function but also proprioceptive and tactile perception for individuals with severe motor impairments. Funded by the Korea Medical Device Development Fund (KMDF), the project spans April 2026 through December 2032 and positions KAIST at the forefront of next‑generation neuro‑rehabilitation technology.
Lead Researchers and Consortium Formation
The effort is co‑led by Professor Kyoungchul Kong, a pioneer in wearable‑robotics and founder of Angel Robotics, and Professor Jung Kim, an expert in robotic skin and somatosensory sensing. Both professors have earned international acclaim—Kong for back‑to‑back Cybathlon gold medals and Kim for the Scientist and Engineer of the Month Award. Their respective teams, together with Angel Robotics Co., Ltd., have formed a consortium that merges expertise in exoskeleton mechanics, neural signal processing, and artificial skin to tackle the ambitious Brain‑to‑Robot challenge.
Technical Approach: Bidirectional Interface
Unlike existing brain‑computer interfaces that focus solely on decoding intent for cursor control or smartphone operation, the Brain‑to‑Robot platform treats the exoskeleton itself as the control target. Brain signals are interpreted to generate movement commands for the robot, while the robot’s sensors capture ground‑reaction force, joint torque, and tactile contact. These sensory streams are then encoded and fed back to the brain via a dedicated neural interface, creating a true bidirectional loop that mirrors the natural sensorimotor pathway.
Components: Exoskeleton Control and Somatosensory Interface
Professor Kong’s team is responsible for the wearable‑robot control system and the AI algorithms that decode movement intention from cortical signals. They are also designing a somatosensory interface—a hardware‑software pipeline that accurately conveys the exoskeleton’s sensory data to the Brain Chip, the semiconductor that processes incoming neural information. This interface must preserve the temporal fidelity of force and tactile cues so that the brain perceives them as genuine bodily sensations.
Robotic Skin and AI Interpretation
Professor Jung Kim’s group is developing robotic skin that can substitute for lost sensation in users with disabilities. The skin incorporates multimodal sensors capable of detecting pressure, vibration, and temperature. AI‑based somatosensory interpretation algorithms translate these raw sensor readings into neural‑compatible patterns that the Brain Chip can inject into the cortex. By mimicking natural afferent signaling, the system aims to restore a realistic sense of touch and limb position.
Signal Processing Challenges and Closed‑Loop Latency
A central technical obstacle is handling hundreds of cortical signal channels while maintaining an ultra‑low‑latency closed loop. The system must decode intent, generate motor commands, acquire sensory feedback, and return it to the brain within milliseconds to preserve the sense of agency and prevent discordant feedback. To achieve this, the teams are advancing high‑bandwidth neural amplifiers, low‑power analog‑to‑digital converters, and real‑time AI encoding/decoding pipelines that operate on dedicated hardware to minimize jitter.
Commercialization Pathway and Angel Robotics Role
Angel Robotics (KOSDAQ: 455900), the spin‑off founded by Professor Kong, will spearhead commercialization. The pathway includes securing regulatory approval from the Ministry of Food and Drug Safety, conducting clinical trials, scaling manufacturing, and deploying the system in rehabilitation centers and eventually home environments. Full‑cycle commercialization seeks to transform the prototype into a medically approved product that can be reimbursed and widely adopted.
Potential Impact on Rehabilitation and Quality of Life
If successful, the Brain‑to‑Robot platform could usher in a new rehabilitation paradigm. Individuals with quadriplegia or severe spinal‑cord injury might regain the ability to walk independently, manipulate objects, and experience tactile feedback at their fingertips during everyday activities. Such recovery would not only improve functional independence but also enhance psychological well‑being by reducing reliance on caregivers and expanding participation in social and vocational life.
Safety, Regulation, and Ethical Considerations
The research consortium emphasizes that long‑term safety, clinical validation, and a robust regulatory framework must evolve alongside the technology. Critical aspects include rigorous efficacy testing, accumulation of clinical evidence, risk‑management systems, protection of brain‑signal data, cybersecurity safeguards, and comprehensive ethical review. Addressing these elements in an integrated fashion is deemed essential for gaining public trust and achieving global market acceptance.
Broader Brain Interface Research at KAIST
Beyond the flagship project, KAIST hosts a spectrum of complementary brain‑interface endeavors. Professor Hyung‑Soon Park’s team studies wearable rehabilitation robots guided by neural intention‑recognition interfaces for neurological disorders. Professor Sungho Cho’s group develops AI‑driven brain‑signal interpretation tools. Professor Jihoon Lee focuses on ultra‑low‑power bio/neural circuits, wireless neural measurement, and on‑device closed‑loop neuromodulation. Professor Hyunjoo Lee’s work centers on high‑resolution multimodal neural electrodes for simultaneous recording and stimulation. Professor Minkyu Je investigates AI‑based semiconductor integrated circuits for neural interfaces, while Professor Jae‑Woong Jeong pursues high‑precision brain‑signal measurement and neuroengineering through neural stimulation.
Conclusion and Vision Statement from KAIST President
KAIST President Kwang‑Hyung Lee praised the Brain‑to‑Robot initiative as a world‑class, highly challenging convergence research effort led by Professors Kong and Kim. He noted that the institute’s extensive expertise in brain interfaces, AI, semiconductors, and robotics provides a solid foundation for pioneering advancements in neuro‑rehabilitation. The ultimate vision, as articulated by Professor Kong, is to enable people with severe motor impairments to move beyond hospital confines, walk unaided, grasp objects, and feel sensations in their daily lives—transforming both capability and quality of life through cutting‑edge brain‑robot symbiosis.