TDL-GNSS
Runzhi Hu
November 11, 2025
A tightly coupled deep learning framework for GNSS positioning in challenging urban environments.
TDL-GNSS is built on top of pyrtklib and TASGNSS, designed to seamlessly integrate deep learning models into the GNSS processing workflow. The framework enables researchers to leverage neural networks for tasks such as satellite signal quality assessment, weight optimization, and positioning error mitigation — all within a unified Python pipeline.
Key Features:
- Tightly integrates deep learning with conventional GNSS processing (SPP, RTK, PPP)
- Built on the established pyrtklib and TASGNSS ecosystem
- End-to-end trainable pipeline for GNSS positioning
- Designed for urban canyon scenarios with severe multipath and NLOS effects
Citation:
@ARTICLE{10965937,
author={Hu, Runzhi and Xu, Penghui and Zhong, Yihan and Wen, Weisong},
journal={IEEE Transactions on Intelligent Transportation Systems},
title={pyrtklib: An Open-Source Package for Tightly Coupled Deep Learning and GNSS Integration for Positioning in Urban Canyons},
year={2025},
volume={26},
number={7},
pages={10652-10662},
doi={10.1109/TITS.2025.3552691}
}