Jesse Morris, Yiduo Wang, and Viorela Ila, “The Importance of Coordinate Frames in Dynamic SLAM.” arXiv, December 2023 [Online]. Available at: http://arxiv.org/abs/2312.04031. [Accessed: January 22, 2024]
@misc{morris_importance_2023,
title = {The {Importance} of {Coordinate} {Frames} in {Dynamic} {SLAM}},
url = {http://arxiv.org/abs/2312.04031},
arxivdoi = {10.48550/arXiv.2312.04031},
freepdf = {https://arxiv.org/pdf/2312.04031.pdf},
urldate = {2024-01-22},
publisher = {arXiv},
author = {Morris, Jesse and Wang, Yiduo and Ila, Viorela},
month = dec,
year = {2023},
note = {arXiv:2312.04031 [cs]},
keywords = {Computer Science - Robotics},
annote = {Comment: 7 pages, 4 figures, submitted to ICRA 2024},
tldr = {A back-end for Dynamic SLAM. Discusses object- versus world-centric dynamic SLAM, advocates for world-centric formulation but evaluates both object- and world-centric versions in factor graph library GTSAM. ``Model free" in that tracked points are used (another option would be object pose). Dynamic objects are assumed to be rigid. An example of where gauge choice leads to different formulations and results.}
}
Most Simultaneous localisation and mapping (SLAM) systems have traditionally assumed a static world, which does not align with real-world scenarios. To enable robots to safely navigate and plan in dynamic environments, it is essential to employ representations capable of handling moving objects. Dynamic SLAM is an emerging field in SLAM research as it improves the overall system accuracy while providing additional estimation of object motions. State-of-the-art literature informs two main formulations for Dynamic SLAM, representing dynamic object points in either the world or object coordinate frame. While expressing object points in a local reference frame may seem intuitive, it may not necessarily lead to the most accurate and robust solutions. This paper conducts and presents a thorough analysis of various Dynamic SLAM formulations, identifying the best approach to address the problem. To this end, we introduce a front-end agnostic framework using GTSAM that can be used to evaluate various Dynamic SLAM formulations.
tl;dr: A back-end for Dynamic SLAM. Discusses object- versus world-centric dynamic SLAM, advocates for world-centric formulation but evaluates both object- and world-centric versions in factor graph library GTSAM. “Model free" in that tracked points are used (another option would be object pose). Dynamic objects are assumed to be rigid. An example of where gauge choice leads to different formulations and results.
The Importance of Coordinate Frames in Dynamic SLAM.
arXiv: http://doi.org/10.48550/arXiv.2312.04031
h/t Amy Tabbpdf: https://arxiv.org/pdf/2312.04031.pdf