Deep Learning-Based Environment Identification for Seamless Localization with GNSS Measurements

Menggang Sheng, Kangwei Wang, Sheng Liu, Xiaoyu Tao, Zhipeng Chen, Zhiqiang Yao

Abstract:	To enhance smartphone positioning accuracy in complex environments, this paper proposes a recognition method based on fine-grained environmental classification. Navigation environments are categorized into seven distinct scenarios: open sky, avenues, semi-outdoor, urban canyon, viaduct-down, shallow indoor, and deep indoor. Data were collected using the smartphone application GnssLogger, yielding a total of 77,616 samples. A classification model was constructed based on the Gated Recurrent Unit (GRU), achieving an accuracy of 99.89% on the training set and 97.31% on a test set comprising 7,530 newly collected samples. The results demonstrate that the proposed method effectively enhances the recognition performance of smartphones in complex environments, providing a foundation for further improvements in positioning accuracy.
Published in:	Proceedings of the 38th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2025) September 8 - 12, 2025 Hilton Baltimore Inner Harbor Baltimore, Maryland
Pages:	1176 - 1186
Cite this article:	Sheng, Menggang, Wang, Kangwei, Liu, Sheng, Tao, Xiaoyu, Chen, Zhipeng, Yao, Zhiqiang, "Deep Learning-Based Environment Identification for Seamless Localization with GNSS Measurements," Proceedings of the 38th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2025), Baltimore, Maryland, September 2025, pp. 1176-1186. https://doi.org/10.33012/2025.20322
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In