This work is a tailored dual-VIO for a mobile manipulator with arm kinematics. The Dual-VIO-MM is an optimization based tightly coupled multi-VIO fusion with arm kinematics. This work contains a baseline dual-VIO (a multi-threaded VINS-Mono), and a proposed dual-VIO-Arm (a multi-threaded VINS-Mono with Arm Odometry).
- For hardware setup, refer to https://github.com/UW-Advanced-Robotics-Lab/lab-wiki/wiki/Waterloo-Steel%3APlatform-Instruction
- Please refer to https://github.com/jaku-jaku/vins-research-pkg
- batch scripts to run over all recorded ros bags:
$HOME/UWARL_catkin_ws/src/waterloo_steel/waterloo_steel_demo/waterloo_steel_analyzer/shortcuts/batch_tmux_vins.sh waterloo_steel_demo_0519 mono_rgb_imu EE d455 all all accurate_T_ic -1 -1 && /home/jx/UWARL_catkin_ws/src/waterloo_steel/waterloo_steel_demo/waterloo_steel_analyzer/shortcuts/batch_tmux_vins.sh waterloo_steel_demo_0519 mono_rgb_imu base d455 all all accurate_T_ic -1 -1
- Public available on ArXiv (Submission to IEEE TMECH): https://arxiv.org/abs/2407.13878
- IEEE TMECH / AIM 2024 Focused Section (Final): https://ieeexplore.ieee.org/document/10542393
@article{Xu_2024,
title={A New Tightly-Coupled Dual-VIO for a Mobile Manipulator With Dynamic Locomotion},
ISSN={1941-014X},
url={http://dx.doi.org/10.1109/TMECH.2024.3400918},
DOI={10.1109/tmech.2024.3400918},
journal={IEEE/ASME Transactions on Mechatronics},
publisher={Institute of Electrical and Electronics Engineers (IEEE)},
author={Xu, Jianxiang and Jeon, Soo},
year={2024},
pages={1–9} }
- This Repo is based on https://github.com/HKUST-Aerial-Robotics/VINS-Fusion
- Uses ceres solver for non-linear optimization and DBoW2 for loop detection, a generic camera model and GeographicLib.
The source code is released under GPLv3 license.