EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Noise-Adversarial Denoising of Motor Acoustic Signals Across Noise Domains

Yang Yang iD, Jiayu Xu, Xiaojun Jiang, Xin Xiao, Hui Dong

Sensors · 15.07.2026 · 26 (14) · 4489

DOI: 10.3390/s26144489 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 15.07.2026. Verfasst von Yang Yang und 4 weiteren Autor:innen. Beteiligte Einrichtung: Quzhou Eco-Industrial Innovation Institute ZJUT, Quzhou 324499, China. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 33 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Quzhou Eco-Industrial Innovation Institute ZJUT.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Motor acoustic monitoring in industrial sites is often compromised by background noise, machine self-noise and shifts between noise domains, which distort the acoustic-signature structure. Here we propose a Noise-Adversarial Denoising Enhancement Network (NADEN) for motor acoustic-signal denoising under complex noise. NADEN maps noisy waveforms into the short-time Fourier transform magnitude domain and predicts a time-frequency mask with an encoder and mask decoder head. This design suppresses noise-dominated time-frequency units while retaining target acoustic structure. A multi-resolution STFT loss constrains spectral consistency across scales, and a gradient reversal layer supports noise-domain adversarial learning. Experiments on a MUSAN-based multi-domain motor acoustic dataset showed that NADEN reduced mean spectral entropy from 0.6288 to 0.6098 while leaving spectral flatness largely unchanged. It also reduced the standard deviation of SFM from 0.0744 to 0.0688, indicating lower spectral randomness and more stable outputs. Sensitivity and ablation experiments support the complementary roles of multi-resolution spectral constraints and noise-domain adversarial learning. These findings suggest a feasible data-driven route for enhancing motor acoustic signals in complex industrial noise environments.

Schlagwörter

Stichwörter der Autor:innen

adversarial learningdenoisingmotor acousticsnoise-domain generalisationtime-frequency masking

Autor:innen und Institutionen

  • Quzhou Eco-Industrial Innovation Institute ZJUT, Quzhou 324499, China.
  • School of Mechanics and Construction Engineering, Jinan University, Guangzhou 510632, China.
  • College of Information Engineering, Zhejiang University of Technology, Hangzhou 310023, China.
  • Zhejiang Key Laboratory of Intelligent Perception and Control for Complex Systems, Zhejiang University of Technology, Hangzhou 310023, China.

ORCID-Profile: Yang Yang

Förderung

  • Quzhou Eco-Industrial Innovation Institute ZJUT

Zitieren

APA

Yang, Y., Xu, J., Jiang, X., Xiao, X., & Dong, H. (2026). Noise-Adversarial Denoising of Motor Acoustic Signals Across Noise Domains. Sensors, 26(14), 4489. https://doi.org/10.3390/s26144489

Vancouver

Yang Y, Xu J, Jiang X, Xiao X, Dong H. Noise-Adversarial Denoising of Motor Acoustic Signals Across Noise Domains. Sensors. 2026;26(14):4489. doi:10.3390/s26144489

BibTeX
@article{Yang2026,
  author = {Yang, Yang and Xu, Jiayu and Jiang, Xiaojun and Xiao, Xin and Dong, Hui},
  title = {Noise-Adversarial Denoising of Motor Acoustic Signals Across Noise Domains},
  journal = {Sensors},
  year = {2026},
  volume = {26},
  number = {14},
  pages = {4489},
  doi = {10.3390/s26144489}
}
Kostenlos starten →