Home

Awesome

R-VIO

R-VIO is an efficient, lightweight, robocentric visual-inertial navigation algorithm for consistent 3D motion tracking using only a monocular camera and a single IMU. Different from the standard world-centric algorithms which directly estimate absolute motion of the mobile platform with respect to a fixed, gravity-aligned, global frame of reference, R-VIO i) estimates relative motion of higher accuracy with respect to a moving, local frame (the IMU frame here), and ii) incrementally updates global pose (orientation and position) through a composition step. This code implements our robocentric sliding-window filtering-based VIO formulation that was originally proposed in our IROS2018 paper and presented in detail in our recent IJRR paper:

@article{huai2022robocentric,
  title={Robocentric visual-inertial odometry},
  author={Huai, Zheng and Huang, Guoquan},
  journal={The International Journal of Robotics Research},
  volume={41},
  number={7},
  pages={667--689},
  year={2022},
  publisher={SAGE Publications Sage UK: London, England}
}

@inproceedings{huai2018robocentric,
  title     = {Robocentric visual-inertial odometry},
  author    = {Huai, Zheng and Huang, Guoquan},
  booktitle = {IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
  pages     = {6319--6326},
  year      = {2018},
  address   = {Madrid, Spain}
}

This work has been further extended in our IEEE RA-L paper below, and the proposed R-VIO2 is also open sourced.

@article{huai2022square,
  title={Square-root robocentric visual-inertial odometry with online spatiotemporal calibration},
  author={Huai, Zheng and Huang, Guoquan},
  journal={IEEE Robotics and Automation Letters},
  volume={7},
  number={4},
  pages={9961--9968},
  year={2022},
  publisher={IEEE}
}

IROS video (EuRoC MAV dataset): YouTube.

IJRR video (9.8km Urban Driving test): YouTube.

1. Prerequisites

We have tested this code under Ubuntu 16.04 and ROS Kinetic.

ROS

Download and install instructions can be found at: http://wiki.ros.org/kinetic/Installation/Ubuntu.

Additional ROS packages needed: tf, sensor_msgs, geometry_msgs, nav_msgs, cv_bridge, eigen_conversions.

Eigen

Download and install instructions can be found at: http://eigen.tuxfamily.org. Required at least 3.1.0.

OpenCV

Download and install instructions can be found at: http://opencv.org. Required at leat 2.4.3. Tested with 2.4.11 and 3.3.1.

2. Build and Run

First, git clone the repository and catkin_make it. Then, to run rvio with single camera/IMU inputs from the ROS topics /camera/image_raw and /imu, a config file in config folder and the corresponding launch file in launch folder (for example, rvio_euroc.yaml and euroc.launch are for EuRoC dataset) are needed, and to visualize the outputs of R-VIO please use rviz with the settings file rvio_rviz.rviz in config folder.

Terminal 1: roscore

Terminal 2: rviz (AND OPEN rvio_rviz.rviz IN THE CONFIG FOLDER)

Terminal 3: roslaunch rvio euroc.launch

Terminal 4: rosbag play --pause V1_01_easy.bag /cam0/image_raw:=/camera/image_raw /imu0:=/imu

Note that when testing the Machine Hall sequences, you should skip the data in the first few seconds (e.g., 40s for MH_01_easy) which are used for initializing the map for SLAM-based algorithms.

You can also run R-VIO with your own sensors (data) by creating a config file rvio_NAME_OF_YOUR_DATA.yaml in config folder and the corresponding launch file NAME_OF_YOUR_DATA.launch in launch folder, referring to our EuRoC example.

3. License

This code is released under GNU General Public License v3 (GPL-3.0).