ExplORB-SLAM

A package for Active visual SLAM using the structure of the underlying pose-graph.

Code used for the paper "ExplORB-SLAM: Active Visual SLAM Exploiting the Pose-graph Topology", accepted for presentation in the Fifth Iberian Robotics Conference (ROBOT 2022).

Tested by jplaced for Ubuntu 20.04, ROS Noetic.

Contact: [email protected], [email protected]

Citation

Placed, J. A., Gómez-Rodríguez, J. J., Tardós, J. D., & Castellanos, J. A. (2022). ExplORB-SLAM: Active Visual SLAM Exploiting the Pose-graph Topology. In 2022 Fifth Iberian Robotics Conference (ROBOT).

Detected dependencies:

• Eigen
• OpenCV
• Python3
  • Numpy
  • Sklearn
  • Numba
  • OpenCV
• Gazebo
• ROS Noetic
  • rviz
  • turtlebot3_teleop
  • gazebo_ros
  • octomap_ros
  • octomap_rviz_plugins
  • move_base

Building

1. Clone repo:

git clone https://github.com/JulioPlaced/ExplORBSLAM.git

2. Build repo:

cd ExplORBSLAM/
catkin b

3. Remember to source the ExplORBSLAM workspace:

source devel/setup.bash

If sourcing doesn't work properly, try

catkin config --no-install
catkin clean --all

and rebuild.

Running

1. Launch the scenario:

AWS house environment:

roslaunch robot_description single_house.launch

Or AWS bookstore environment:

roslaunch robot_description single_bookstore.launch

2. Launch the decision maker

roslaunch decision_maker autonomous_agent.launch