[Edaily Reporter Shin Yeong-bin ] ANGEL ROBOTICS Co., Ltd.(455900)has unveiled a wearable robot platform for physical artificial intelligence (AI) research. The platform enables researchers to collect human motion data and test new AI and control algorithms without having to build the robot hardware themselves.

ANGEL ROBOTICS Co., Ltd. announced on the 29th that it has unveiled “phai-x1,” a wearable robot designed for physical AI research and development.

The phai-x1 was developed based on ANGEL ROBOTICS Co., Ltd.’s gait-assistance robot, the “Angel Suit H10.” It is designed to allow researchers in fields such as robotics, control engineering, physical AI, biomechanics, human motion analysis, and biomedical and rehabilitation engineering to develop and test their own algorithms on a common hardware platform.

A key feature is that it reduces the burden on researchers of having to build a separate wearable robot from scratch to suit their research objectives. Researchers can configure their desired programs using expansion modules responsible for implementing user algorithms and collecting data; they can also execute the developed control algorithms at a rate of 1,000 times per second to immediately apply and validate them on the actual robot.

It also features the ability to simultaneously collect data on both human and robot movements. The Pi-X1 independently controls the left and right hip joints and records the robot’s joint movements and status data.

Additionally, users can connect an inertial measurement unit (IMU) to measure body tilt and movement, an electromyography (EMG) sensor to monitor muscle activity, and a ground reaction force (GRF) sensor to measure the force generated when the foot strikes the ground. According to the developers, this allows both human movement and the robot’s responses to be utilized together as data for AI training.

The collected data can be used for research ranging from the analysis of human gait and human-robot interaction to the estimation of movement intent, user-customized assistive control, and AI reinforcement learning-based control.

The Pi-X1 was used as a research platform at the “IFAC Wearable Robotics Challenge 2026,” held last August. According to the company, after receiving approximately two hours of training, participants spent two days implementing various algorithms, including game interfaces using wearable robots, control for walking on uneven terrain, slip detection, support for hemiplegic walkers, and EMG-based performance evaluation.

ANGEL ROBOTICS Co., Ltd. plans to supply the Pi-X1 to universities, public research institutions, corporate research labs, university hospital and medical school laboratories, and biomedical and rehabilitation engineering research institutions. The company aims to expand the product’s scope of application beyond sales to include educational programs, research projects, competitions, and joint research initiatives.

Kong Kyung-chul, Director of the Future Technology Institute at ANGEL ROBOTICS Co., Ltd., stated, “Camera footage alone is not sufficient for AI to properly understand human movement,” adding, “The Pi-X1 is a platform that records and stores as data the forces exchanged between humans and robots, the timing of muscle activation, and the forces generated when the foot makes contact with the ground.”