IGLOV

Active View Planning for Visual SLAM in Outdoor Environments Based on Continuous Information Modeling

1. Introduction

IGLOV is a approach by actively and smoothly controlling a gimbal camera equipped on the robot to realize robust and accurate SLAM in unknown outdoor environments. The main works include three-fold:

• Feature distribution-weighted Fisher information to represent the localization uncertainty of environments.
• A continuous information modeling method to map the environmental information, like localization uncertainty, into multiple polynomial functions.
• An information gradient-based local view (IGLOV) planner plans the optimal camera views in real-time for obtaining maximal environmental information.

Authors: Zhihao Wang, Haoyao Chen from the Networked RObotics and Sytems Lab, Shiwu Zhang, and Yunjiang Lou.

If you use IGLOV for your academic research, please cite the following paper [pdf].

@ARTICLE{10129820,
  author={Wang, Zhihao and Chen, Haoyao and Zhang, Shiwu and Lou, Yunjiang},
  journal={IEEE/ASME Transactions on Mechatronics}, 
  title={Active View Planning for Visual SLAM in Outdoor Environments Based on Continuous Information Modeling}, 
  year={2023},
  volume={},
  number={},
  pages={1-12},
  doi={10.1109/TMECH.2023.3272910}}

1.1. Framework

1.2. Related Video:

Video links: Youtube or Bilibili.

2. Installation

Tested on Ubuntu 16.04 and 18.04.

• Install ROS packages

sudo apt-get install ros-(rosversion)−octomap−rosros−(rosversion)-octomap-ros ros-(rosversion)-octomap-msgs ros-(rosversion)−yocs−cmd−vel−muxros−(rosversion)-yocs-cmd-vel-mux ros-(rosversion)-pcl-ros

cd {YOUR_PATH}
git clone https://github.com/HITSZ-NRSL/IGLOV.git
cd IGLOV
catkin build

• Add Gazebo model path

export GAZEBO_MODEL_PATH=YOURPATHYOUR_PATH/IGLOV/src/plt1000_base/plt_gazebo/models:$GAZEBO_MODEL_PATH

3. Run examples

Gazebo

cd {YOUR_PATH}/IGLOV
roslaunch plt_gazebo gazebo_realsense.launch

SLAM

cd {YOUR_PATH}/IGLOV
roslaunch orb_slam2_ros orb_slam2_d435_rgbd_realsense.launch

View Planning

cd {YOUR_PATH}/IGLOV
roslaunch view_planning view_planning.launch

Robot Base

cd {YOUR_PATH}/IGLOV/src/plt1000_base/plt_control/src
python robot_keyboard_teleop.py 

Then press the key 'r'(lowercase), the robot base will follow the preset trajectory.

5. Acknowledgments

SLAM in IGLOV is based on ORB-SLAM2.

LICENSE

The source code is released under GPLv2 license.