Adaptive Convolutional Neural Network-Enhanced Scale-Fusion Network for Human Activity Recognition Using Wearable Sensors
Sensors · 2026
DOI: 10.3390/s26185879 ↗Kurz gefasst
Fachartikel in der Zeitschrift Sensors, 2026. Verfasst von Dongpeng Xie und 4 weiteren Autor:innen. Beteiligte Einrichtung: GNSS Research Center, Wuhan University, Wuhan 430079, China. Hauptthemen laut MEDLINE: Wearable Electronic Devices, Human Activities. Die Arbeit nennt 44 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Key R&D Program of China, Hubei Luojia Laboratory.
Abstract (Originalsprache)
Deep learning has achieved notable success in sensor-based human activity recognition (SHAR), yet wearable inertial signals contain activity patterns at different temporal scales while practical deployment imposes strict computational constraints. This paper proposes a three-stage Adaptive CNN-Enhanced Scale-fusion Network (ACESNet) for six-channel accelerometer–gyroscope activity recognition. The Adaptive Kernel Encoder dynamically combines temporal kernels of different sizes, while the CNN-Enhanced Scale-fusion design combines multi-scale temporal mixing, explicit channel mixing and a parallel local CNN path without materializing a global attention matrix. The main benchmark uses a fixed stratified 70/15/15 window-level partition (split seed 42), and all methods in the main comparison are repeated over five training seeds. On REALWORLD, MotionSense, UCI-HAR and SHL, ACESNet achieves accuracies of 95.37 ± 0.09%, 99.33 ± 0.12%, 98.01 ± 0.23% and 93.49 ± 0.10%, with Macro-F1 scores of 95.59 ± 0.09%, 99.12 ± 0.15%, 98.16 ± 0.21% and 94.03 ± 0.08%, respectively. A separate official subject-independent UCI-HAR evaluation yields 94.01 ± 0.84% accuracy and 94.11 ± 0.88% Macro-F1. The adaptive-kernel selector has a mean normalized entropy of 0.9114, providing evidence against severe single-branch collapse rather than strong kernel selectivity. On a Raspberry Pi 5 (Raspberry Pi Ltd., Cambridge, UK), ACESNet requires 8.731 MFLOPs and 1.606 ± 0.014 ms per window in FP32 ONNX Runtime 1.18.1. These results support a favorable performance–efficiency trade-off relative to the selected baselines under the stated protocol.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- GNSS Research Center, Wuhan University, Wuhan 430079, China.
- Hubei Luojia Laboratory, Wuhan 430079, China.
- University Student Engineering Training and Innovation Practice Center, Wuhan University, Wuhan 430072, China.
ORCID-Profile: Peihui Yan
Förderung
- National Key R&D Program of China
- Hubei Luojia Laboratory
Zitieren
APA
Xie, D., Yan, P., Li, Y., Wang, Q., & Liu, J. (2026). Adaptive Convolutional Neural Network-Enhanced Scale-Fusion Network for Human Activity Recognition Using Wearable Sensors. Sensors. https://doi.org/10.3390/s26185879
Vancouver
Xie D, Yan P, Li Y, Wang Q, Liu J. Adaptive Convolutional Neural Network-Enhanced Scale-Fusion Network for Human Activity Recognition Using Wearable Sensors. Sensors. 2026. doi:10.3390/s26185879
BibTeX
@article{Xie2026,
author = {Xie, Dongpeng and Yan, Peihui and Li, Yifei and Wang, Qinghai and Liu, Jingnan},
title = {Adaptive Convolutional Neural Network-Enhanced Scale-Fusion Network for Human Activity Recognition Using Wearable Sensors},
journal = {Sensors},
year = {2026},
doi = {10.3390/s26185879}
}
EUVIMED