Towards End-to-End GPS Localization with Neural Pseudorange Correction (2401.10685v2)
Abstract: The pseudorange error is one of the root causes of localization inaccuracy in GPS. Previous data-driven methods regress and eliminate pseudorange errors using handcrafted intermediate labels. Unlike them, we propose an end-to-end GPS localization framework, E2E-PrNet, to train a neural network for pseudorange correction (PrNet) directly using the final task loss calculated with the ground truth of GPS receiver states. The gradients of the loss with respect to learnable parameters are backpropagated through a Differentiable Nonlinear Least Squares (DNLS) optimizer to PrNet. The feasibility of fusing the data-driven neural network and the model-based DNLS module is verified with GPS data collected by Android phones, showing that E2E-PrNet outperforms the baseline weighted least squares method and the state-of-the-art end-to-end data-driven approach. Finally, we discuss the explainability of E2E-PrNet.
- Improving GPS code phase positioning accuracy in urban environments using machine learning. IEEE Internet of Things Journal, 8(8):7065–7078, 2020.
- Prediction on the urban GNSS measurement uncertainty based on deep learning networks with long short-term memory. IEEE Sensors Journal, 21(18):20563–20577, 2021.
- PrNet: A neural network for correcting pseudoranges to improve positioning with android raw GNSS measurements. arXiv preprint arXiv:2309.12204, 2023.
- Improving GNSS positioning using neural-network-based corrections. NAVIGATION: Journal of the Institute of Navigation, 69(4), 2022.
- Neural moving horizon estimation for robust flight control. IEEE Transactions on Robotics, 2023.
- Differentiable integrated motion prediction and planning with learnable cost function for autonomous driving. IEEE Transactions on Neural Networks and Learning Systems, 2023.
- Differentiable factor graph optimization with intelligent covariance adaptation for accurate smartphone positioning. In Proceedings of the 36th International Technical Meeting of the Satellite Division of The Institute of Navigation, pages 2765–2773, 2023.
- Theseus: A library for differentiable nonlinear optimization. Advances in Neural Information Processing Systems, 35:3801–3818, 2022.
- Android raw GNSS measurement datasets for precise positioning. In Proceedings of the 33rd international technical meeting of the satellite division of the Institute of Navigation, pages 1925–1937, 2020.
- Xu Weng and Keck Voon Ling. Localization with noisy android raw gnss measurements. In 2023 IEEE Asia Pacific Conference on Wireless and Mobile (APWiMob), pages 95–101. IEEE, 2023.
- Deep learning is robust to massive label noise. arXiv preprint arXiv:1705.10694, 2017.