TC-VIML

Xi Zheng November 11, 2025

Tightly-coupled Visual/Inertial/Map Integration with Observability Analysis for Reliable Localization of Intelligent Vehicles

To enable reliable and drift-free localization for intelligent vehicles, we propose a tightly-coupled visual-inertial odometry (VIO) system that leverages a 3D prior line map. Unlike loosely-coupled methods, our approach deeply integrates line features into a factor graph optimization, supported by a robust cross-modality matching and outlier rejection strategy. For the first time, we rigorously prove that our system achieves full observability in global translation, leaving only the yaw angle unobservable. Evaluations in both simulated and real-world environments confirm the system’s effectiveness.

For details, please refer to our official repository at TC-VIML

And if you are using this code, please cite our paper by

@article{zheng2024tightly,
  title={Tightly-coupled visual/inertial/map integration with observability analysis for reliable localization of intelligent vehicles},
  author={Zheng, Xi and Wen, Weisong and Hsu, Li-Ta},
  journal={IEEE Transactions on Intelligent Vehicles},
  year={2024},
  publisher={IEEE}
}

@inproceedings{zheng2023tightly,
  title={Tightly-coupled line feature-aided visual inertial localization within lightweight 3d prior map for intelligent vehicles},
  author={Zheng, Xi and Wen, Weisong and Hsu, Li-Ta},
  booktitle={2023 IEEE 26th International Conference on Intelligent Transportation Systems (ITSC)},
  pages={6019--6026},
  year={2023},
  organization={IEEE}
}