Optimizing wearable IMU configurations for running gait analysis: a machine learning-based sensor fusion approach
Frontiers in Bioengineering and Biotechnology · 11.02.2026 · 14
DOI: 10.3389/fbioe.2026.1762919 ↗Kurz gefasst
Fachartikel in der Zeitschrift Frontiers in Bioengineering and Biotechnology, online erschienen am 11.02.2026. Verfasst von Ye Yuan und 3 weiteren Autor:innen. Beteiligte Einrichtung: College of Physical Education, Xuzhou University of Technology, Xuzhou, Jiangsu, China. Die Arbeit wurde bisher 6-mal zitiert (Crossref) und nennt 29 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Objective
This study applies machine learning (ML) techniques to address this hardware limitation by determining the feasibility of reducing a high-dimensional 17-sensor network to a “minimal-optimal” subset without compromising measurement accuracy. Unlike previous studies focusing on activity classification, we systematically quantify the information redundancy in kinematic chains to optimize sensor fusion architectures.
Methods
Twenty-five recreational runners performed treadmill protocols at three speeds (8, 10, and 12 km/h) while wearing a gold-standard Xsens MVN system (17 IMUs). Raw accelerometer and gyroscope signals were programmatically subsetted to simulate minimal configurations. A Random Forest (RF) regression model was selected after benchmarking against baseline Linear Regression and deep learning (LSTM) models. A comprehensive vector of time-and frequency-domain features was extracted via sliding windows, and Recursive Feature Elimination (RFE) was applied to identify the most critical signal attributes.
Results
Analysis revealed that a single lumbosacral IMU could successfully reconstruct global parameters (Cadence, Vertical Oscillation, Ground Contact Time) with high precision ( R 2 > 0.95 , M A P E < 5 % ), outperforming standard commercial benchmarks. However, this single-node setup failed to detect gait asymmetry ( R 2 = 0.52 ). A distributed three-sensor fusion configuration (Lumbosacral + Bilateral Ankles) resolved this limitation, achieving results comparable to the full-body system for all parameters ( R 2 > 0.91 , M A P E = 7.12 % ). Performance remained robust across all running speeds, with only a marginal accuracy drop at 12 km/h.
Conclusion
This study validates a machine learning framework for optimizing sensor array design. The proposed three-sensor fusion offers a robust, low-cost architectural blueprint for next-generation wearable devices, proving that complex deep learning is not always required when sensor placement is biomechanically optimized.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- College of Physical Education, Xuzhou University of Technology, Xuzhou, Jiangsu, China.
- Graduate School, Tianjin University of Traditional Chinese Medicine, Tianjin, China.
- Faculty of Health and Medicine, Lancaster University, Lancaster, United Kingdom.
- General Education College, Cyberspace Security University of China, Wuhan, China.
Zitieren
APA
Yuan, Y., Yu, Y., Cai, S., & Cheng, W. (2026). Optimizing wearable IMU configurations for running gait analysis: a machine learning-based sensor fusion approach. Frontiers in Bioengineering and Biotechnology, 14. https://doi.org/10.3389/fbioe.2026.1762919
Vancouver
Yuan Y, Yu Y, Cai S, Cheng W. Optimizing wearable IMU configurations for running gait analysis: a machine learning-based sensor fusion approach. Frontiers in Bioengineering and Biotechnology. 2026;14. doi:10.3389/fbioe.2026.1762919
BibTeX
@article{Yuan2026,
author = {Yuan, Ye and Yu, Yaohui and Cai, Shanshan and Cheng, Weidong},
title = {Optimizing wearable IMU configurations for running gait analysis: a machine learning-based sensor fusion approach},
journal = {Frontiers in Bioengineering and Biotechnology},
year = {2026},
volume = {14},
doi = {10.3389/fbioe.2026.1762919}
}
EUVIMED