EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Domain-Adaptive Direction of Arrival (DOA) Estimation in Complex Indoor Environments Based on Convolutional Autoencoder and Transfer Learning

Lingyu Shen, Jianfeng Li, Jingjing Pan, Junpeng Shi, Rui Xu, Hao Wang, Weiming Deng iD

Sensors · 08.05.2025 · 25 (10) · 2959

DOI: 10.3390/s25102959 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 08.05.2025. Verfasst von Lingyu Shen und 6 weiteren Autor:innen. Beteiligte Einrichtung: College of Electronic and Information Engineering, Nanjing University of Aeronautics and Asteonautics, Nanjing 211106, China. Die Arbeit wurde bisher 6-mal zitiert (Crossref) und nennt 41 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: the National Science Foundation of China, the Key R&D Plan of Jiangsu Province, the China Postdoctoral Science Foundation.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Direction of arrival (DOA) estimation for signal sources in indoor environments has become increasingly important in wireless communications and smart home applications. However, complex indoor conditions, such as multipath effects and noise interference, pose significant challenges to estimation accuracy. This issue is further complicated by domain discrepancies in data collected from different environments. To address these challenges, we propose a deep domain-adaptation-based DOA estimation method. The approach begins with deep feature extraction using a Convolutional Autoencoder (CAE) and employs a Domain-Adversarial Neural Network (DANN) for domain adaptation. By integrating Gradient Reversal Layer (GRL) and Maximum Mean Discrepancy (MMD) loss functions, the model effectively reduces distributional differences between the source and target domains. The CAE-DANN enables transfer learning between data with similar features from different domains. With minimal labeled data from the target domain incorporated into the source domain, the model leverages labeled source data to adapt to unlabeled target data. GRL counters domain shifts, while MMD refines feature alignment. Experimental results show that, in complex indoor environments, the proposed method outperforms other methods in terms of overall DOA prediction performance in both the source and target domains. This highlights a robust and practical solution for high-precision DOA estimation in new environments, requiring minimal labeled data.

Schlagwörter

Stichwörter der Autor:innen

DOADeep Learningdomain adaptationindoor scenes

Autor:innen und Institutionen

  • College of Electronic and Information Engineering, Nanjing University of Aeronautics and Asteonautics, Nanjing 211106, China.

ORCID-Profile: Weiming Deng

Förderung

  • the National Science Foundation of China
  • the Key R&D Plan of Jiangsu Province
  • the China Postdoctoral Science Foundation
  • the China University Innovation Fund for Production, Education and Research
  • National Science Foundation of China
  • Key R&D Plan of Jiangsu Province
  • China Postdoctoral Science Foundation
  • China University Innovation Fund for Production, Education and Research

Zitieren

APA

Shen, L., Li, J., Pan, J., Shi, J., Xu, R., Wang, H., & Deng, W. (2025). Domain-Adaptive Direction of Arrival (DOA) Estimation in Complex Indoor Environments Based on Convolutional Autoencoder and Transfer Learning. Sensors, 25(10), 2959. https://doi.org/10.3390/s25102959

Vancouver

Shen L, Li J, Pan J, Shi J, Xu R, Wang H, et al. Domain-Adaptive Direction of Arrival (DOA) Estimation in Complex Indoor Environments Based on Convolutional Autoencoder and Transfer Learning. Sensors. 2025;25(10):2959. doi:10.3390/s25102959

BibTeX
@article{Shen2025,
  author = {Shen, Lingyu and Li, Jianfeng and Pan, Jingjing and Shi, Junpeng and Xu, Rui and Wang, Hao and Deng, Weiming},
  title = {Domain-Adaptive Direction of Arrival (DOA) Estimation in Complex Indoor Environments Based on Convolutional Autoencoder and Transfer Learning},
  journal = {Sensors},
  year = {2025},
  volume = {25},
  number = {10},
  pages = {2959},
  doi = {10.3390/s25102959}
}
Kostenlos starten →