Hyperion – A Fast, Versatile Symbolic Gaussian Belief Propagation Framework for Continuous-Time SLAM
Abstract
CTSLAM has become a promising approach for fusing asynchronous and multi-modal sensor suites. Unlike discrete-time SLAM, which estimates poses discretely, CTSLAM uses continuous-time motion parametrizations, facilitating the integration of a variety of sensors such as rolling-shutter cameras, event cameras and IMU. However, CTSLAM approaches remain computationally demanding and are conventionally posed as centralized NLLS optimizations. Targeting these limitations, we not only present the fastest SymForce-based [?] B- and Z-Spline implementations achieving speedups between 2.43x and 110.31x over Sommer [?] but also implement a novel continuous-time GBP framework, coined , which targets decentralized probabilistic inference across agents. We demonstrate the efficacy of our method in motion tracking and localization settings, complemented by empirical ablation studies. Code: https://github.com/VIS4ROB-lab/hyperion
Cite
Text
Hug et al. "Hyperion – A Fast, Versatile Symbolic Gaussian Belief Propagation Framework for Continuous-Time SLAM." Proceedings of the European Conference on Computer Vision (ECCV), 2024. doi:10.1007/978-3-031-73404-5_13Markdown
[Hug et al. "Hyperion – A Fast, Versatile Symbolic Gaussian Belief Propagation Framework for Continuous-Time SLAM." Proceedings of the European Conference on Computer Vision (ECCV), 2024.](https://mlanthology.org/eccv/2024/hug2024eccv-hyperion/) doi:10.1007/978-3-031-73404-5_13BibTeX
@inproceedings{hug2024eccv-hyperion,
title = {{Hyperion – A Fast, Versatile Symbolic Gaussian Belief Propagation Framework for Continuous-Time SLAM}},
author = {Hug, David and Alzugaray, Ignacio and Chli, Margarita},
booktitle = {Proceedings of the European Conference on Computer Vision (ECCV)},
year = {2024},
doi = {10.1007/978-3-031-73404-5_13},
url = {https://mlanthology.org/eccv/2024/hug2024eccv-hyperion/}
}