Knee Angle-Based Lower-Limb Locomotion Detection Algorithm With Sensor-Minimal Design
IEEE Transactions on Automation Science and Engineering · 2026 · 23 · 13744-13756
DOI: 10.1109/tase.2026.3717331 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Automation Science and Engineering, 2026. Verfasst von Hyun Jun Cho und 5 weiteren Autor:innen. Die Arbeit nennt 39 Literaturangaben. Gefördert von: Research and Development Program for Forest Science Technology provided by Korea Forest Service, “Development of Textile Materials Applied to Wearable Robot Suits to Improve the Combat Power of Soldiers” provided by the Ministry of Trade, Industry & Energy…, Council of Science & Technology (NST) grant by Korean Government [Ministry of Science and Information and Communication Technology (MSIT)].
Abstract (Originalsprache)
This study proposes an algorithm for real-time recognition of five locomotion modes—walking, running, squatting, standing posture, and sitting posture—using only knee angle data to control a lower-limb exoskeleton. Physically demanding tasks in various occupational settings impose substantial strain on joints and muscles, underscoring the necessity of assistive exoskeletons to mitigate fatigue and reduce the risk of musculoskeletal injuries. Since the lower body bears the majority of the load during repetitive movements, robotic assistance is crucial for alleviating stress on the knees and surrounding muscle groups. Accurate identification of key locomotion modes is essential for seamless human-robot interaction and optimal performance. To enhance user comfort, the system is designed with a minimal hardware configuration, relying solely on the actuator’s encoder to measure knee angle and angular velocity. The proposed algorithm accurately detects user motion by classifying each leg’s state and subsequently merging the states of both legs to recognize five locomotion modes: walking, running, sitting posture, standing posture, and squatting. Experimental validation was conducted with seven healthy male participants performing various locomotion tasks on a treadmill and overground, demonstrating that the exoskeleton can accurately recognize user movements without additional sensors and indicating the potential applicability of the proposed method. Note to Practitioners—This study was motivated by the need to reduce the excessive load on the knees during repetitive or heavy-lifting tasks commonly encountered in industrial, military, and other physically demanding work environments. Existing motion detection systems typically rely on multiple sensors, such as IMUs, force sensors, or external motion capture systems, to improve accuracy. However, this increases hardware complexity, raises maintenance costs, and can lead to discomfort for users due to the additional weight and wiring. To address these issues, this paper proposes a real-time motion detection method that identifies five key movements—walking, running, squatting, standing, and sitting—using only knee angle and angular velocity data estimated from a single actuator encoder. By eliminating the need for multiple sensors, this approach simplifies the overall system, reduces mechanical interference, and enhances wearability, making it more practical for long-term use in industrial and military applications. Despite these advantages, the proposed method has some limitations. In particular, it requires further validation for handling abrupt motion transitions and non-standard postures, which could affect classification accuracy in dynamic environments. Future research will focus on refining the algorithm to improve robustness under diverse movement patterns and integrating it with assistive torque control strategies to provide real-time strength augmentation. Ultimately, this work aims to contribute to the development of more effective and user-friendly wearable assistive devices that improve worker safety and efficiency while minimizing system complexity.
Förderung
- Research and Development Program for Forest Science Technology provided by Korea Forest Service
- “Development of Textile Materials Applied to Wearable Robot Suits to Improve the Combat Power of Soldiers” provided by the Ministry of Trade, Industry & Energy…
- Council of Science & Technology (NST) grant by Korean Government [Ministry of Science and Information and Communication Technology (MSIT)]
Zitieren
APA
Cho, H. J., Yoon, S., Song, J., Han, S., Choi, H. R., & Cho, J. (2026). Knee Angle-Based Lower-Limb Locomotion Detection Algorithm With Sensor-Minimal Design. IEEE Transactions on Automation Science and Engineering, 23, 13744-13756. https://doi.org/10.1109/tase.2026.3717331
Vancouver
Cho HJ, Yoon S, Song J, Han S, Choi HR, Cho J. Knee Angle-Based Lower-Limb Locomotion Detection Algorithm With Sensor-Minimal Design. IEEE Transactions on Automation Science and Engineering. 2026;23:13744-13756. doi:10.1109/tase.2026.3717331
BibTeX
@article{Cho2026,
author = {Cho, Hyun Jun and Yoon, Sungwoon and Song, Junyong and Han, Sangchul and Choi, Hyouk Ryeol and Cho, Jungsan},
title = {Knee Angle-Based Lower-Limb Locomotion Detection Algorithm With Sensor-Minimal Design},
journal = {IEEE Transactions on Automation Science and Engineering},
year = {2026},
volume = {23},
pages = {13744--13756},
doi = {10.1109/tase.2026.3717331}
}
EUVIMED