CES 2024 Innovation Awards Honoree
H-FlexTailored to users, Customizable wearable robots
Category Robotics
View on CES.tech
Author: Saikat Sahoo†, Yoosun Kim†, Dongmin Go, Sejun Park, Jeongho Choo, Gyeongmo Kim, Cheonkyu Park, Giuk Lee
Corresponding author: Giuk Lee
Journal: Biomimetics
Published date: 2026.11
Link: https://www.mdpi.com/2313-7673/11/11/782
DOI: 10.3390/biomimetics11110782
본 연구는 엑소슈트의 보조력을 사용자에게 맞추는 Human-in-the-Loop(HITL) 최적화를 호흡 가스 분석 장비 없이 엑소슈트에 달린 센서만으로 수행할 수 있는지 확인하기 위해 진행되었습니다. 기존 HITL 최적화는 대사 에너지를 재는 호흡 가스 분석 장비가 필요해 실험실 밖에서 쓰기 어렵고, 이를 대신하는 지표들은 센서를 추가로 달아야 하거나 대사 에너지와의 관계가 충분히 확인되지 않았습니다. 연구진은 이전 연구에서 목표 함수로 사용한 양의 기계적 일률(positive mechanical power, PMP)에 더해, 허벅지에 부착된 관성 센서(IMU)로 측정하는 보행 주기 시간(stride time, ST)이 대사 에너지 감소를 더 잘 나타내는지 살펴보았습니다.
참가자 3명이 힙 신전 보조 엑소슈트를 착용하고 5° 경사의 트레드밀을 1 m/s로 걸은 그리드 서치 데이터(6건)를 분석한 결과, 보행 주기 시간과 대사 에너지 감소의 상관계수는 0.63 ± 0.14로 PMP(0.51 ± 0.37)보다 높았습니다. 두 지표를 다중 선형 회귀로 결합한 다변수 비용 함수(multi-variable cost function, MVCF)는 상관계수 0.68 ± 0.11로 가장 높았고 평균 절대 오차도 가장 낮았습니다. 다만 비용 함수 간 차이는 통계적으로 유의하지 않았습니다. 이어 참가자 5명을 대상으로 MVCF를 목표 함수로 한 베이지안 HITL 최적화를 진행한 결과, 최적화된 보조 조건에서 순 대사율이 전원을 끈 엑소슈트 착용 조건 대비 13.3%, 엑소슈트 미착용 대비 7.5%, 생체 모사 보조 프로파일 대비 5.2% 감소했습니다.
연구진은 장치가 측정하는 기계적 일률과 사람의 보행 타이밍을 함께 활용하면, 대사 측정 장비 없이 엑소슈트 자체 센서만으로 보조력을 개인에게 맞출 수 있는 비용 함수를 만들 수 있다고 결론 내렸습니다. 이번 결과는 트레드밀 경사 보행에서 소수의 참가자를 대상으로 얻은 탐색적 결과로, 다른 보행 환경에서의 추가 검증이 필요합니다.
This study examined whether human-in-the-loop (HITL) optimization, which personalizes exosuit assistance to each user, can be performed using only the sensors already built into the exosuit, without respirometry equipment. Conventional HITL optimization relies on bulky equipment to measure metabolic cost, and existing proxy measures either require additional sensors or have an unquantified relationship with metabolism. Building on their previous approach based on positive mechanical power (PMP), the researchers investigated whether stride time (ST), measured by the exosuit's thigh-mounted inertial sensors, better represents metabolic reduction.
Using grid-search data from three participants walking at 1 m/s on a 5° incline with a hip extension exosuit (six subject-day cases), stride time correlated with metabolic reduction numerically better than PMP (r = 0.63 ± 0.14 vs. 0.51 ± 0.37). A multi-variable cost function (MVCF) combining the two by multiple linear regression performed best (r = 0.68 ± 0.11) with the lowest mean absolute error, although the differences between cost functions were not statistically significant. When the MVCF was used as the objective of Bayesian HITL optimization in five participants, the optimized assistance reduced net metabolic rate by 13.3%, 7.5%, and 5.2% compared with the power-off, no-suit, and biomimetic conditions, respectively.
The researchers concluded that combining device-side mechanical power with human-side gait timing may provide an onboard-sensor cost function grounded in measured metabolic data. These are exploratory results from treadmill incline walking with a small number of participants, and further validation in other walking conditions is needed.
Funding information:
National Research Foundation of Korea (NRF) grants funded by the Korea government (MSIT)
CES 2024 Innovation Awards Honoree
H-FlexTailored to users, Customizable wearable robots
Category Robotics
View on CES.tech
CES 2025 Innovation Awards Honoree
H-MediAll-in-one Gait Rehabilitation and Prognosis
Category Accessibility & Longevity, Robotics
View on CES.tech
CES 2026 Innovation Awards Honoree
H-Medi ProComfortable Personalized Robotic Suit for Targeted Recovery
Category Robotics
View on CES.tech[04315] HUROTICS Inc., 13F, Yongsan The Prime, 11, Wonhyo-ro 90-gil, Yongsan-gu, Seoul, Republic of Korea
Hurotics Inc. (hereinafter referred to as the "Company") establishes and discloses the following personal information processing guidelines to protect the personal information of data subjects and handle related grievances promptly and smoothly in accordance with Article 30 of the Personal Information Protection Act.
The Company processes personal information for the following purposes. The personal information being processed will not be used for purposes other than the following, and if the purpose of use changes, necessary measures will be taken, such as obtaining separate consent in accordance with Article 18 of the Personal Information Protection Act.
- Within 3 years from the date of receipt of website customer inquiries
- Request to view personal information
- Request for correction in case of errors
- Request for deletion
- Request to stop processing
- Required items: Name, Contact information (phone number, email address), Company name
- IP address, cookies, MAC address, service usage records, visit records, malicious use records, etc.
- Destruction procedure
The Company selects personal information for which destruction reasons have occurred and destroys it with the approval of the Company's personal information protection officer.
- Destruction method
Personal information printed on paper is destroyed by shredding or incineration. Personal information stored in electronic file format is deleted using technical methods that prevent recovery of the records.
Posted: December 19, 2025