Wearable AI Helps Parkinson’s Patients Take Their First Steps: Challenge by Three Hanyang University Undergraduates

[Financial News] A wearable artificial intelligence (AI) device that helps prevent patients with Parkinson’s disease from falling when their feet suddenly freeze while walking has been developed by university undergraduates. The small device worn around the ankle analyzes gait on its own and detects risks in advance, without connecting to an external server or communications network.
Hanyang University announced on the 11th that Hanggeoreum, a team comprising third-year students Kang Yu-ri, Kwon Dae-yoon and Jin Ye-rim from the Department of Electronic Engineering, developed a low-power wearable system to support safe walking for patients with Parkinson’s disease and received the Korea Information Science Promotion Association Chairman’s Award in the university division at ICT AWARD KOREA 2026.
The students focused on freezing of gait, a common symptom among patients with Parkinson’s disease. The condition causes the feet to freeze as if stuck to the ground while walking or turning. Because the upper body continues moving forward, patients can lose their balance and fall heavily, making them vulnerable to injury. Fear of falling leads many patients to avoid going outside and remain at home, creating a vicious cycle that reduces their quality of life.
Hanggeoreum focused on creating a device that patients could wear throughout their daily lives without discomfort. Previous studies often used multiple sensors attached to different parts of the body or complex AI programs, making the devices heavy and causing severe battery drain.
The students found a solution by combining a single inertial measurement unit (IMU) sensor with an ultra-lightweight machine-learning algorithm that requires little computing power. The gait data collected by the sensor is analyzed in real time by a microcontroller unit (MCU) inside the device, without being sent to a smartphone or internet cloud. By adopting an on-device AI approach, in which AI operates directly on the device, the team significantly reduced power consumption and response delays.
The device finely classifies and recognizes the user’s movements in five stages: stationary, normal walking, pre-freezing, signal transmission and recovery. It also incorporates an intelligent power-saving function that reduces data collection when movement is limited and operates intensively only when abnormal signs appear.
The achievement was highly regarded because the third-year undergraduates completed the entire process themselves, from hardware circuit production to software coding, rather than stopping at the idea stage. Kwon Dae-yoon designed and produced the printed circuit board (PCB), Kang Yu-ri handled AI model compression and the internal control program, and Jin Ye-rim was responsible for data analysis and performance verification, resulting in a working device.
The current prototype uses an LED to indicate the user’s status when it detects signs preceding freezing of gait. The team is further developing a haptic feedback function that delivers immediate vibrations to the ankle to help patients move a frozen foot again. In the future, the team plans to connect the device with smartphones or smartwatches so that family members and medical professionals can monitor changes in patients’ conditions together.
Kwon Dae-yoon said, "Freezing of gait is a major problem that threatens patients’ safety in daily life and even undermines their will to live." He added, "I want to complete a system that helps patients take a step through vibration signals at dangerous moments, so they can walk out into the world without fear."  

[email protected] Kim Man-gi Reporter