HFSD

History-Aware Free Space Detection for Efficient Autonomous Exploration using Aerial Robots in ROS

Overview

This package is capable of proposing appropriate directions for exploration by utilizing a sliding-window history of the robot’s pose estimates and the depth measurements of the environment to identify the directions of probable unobservable free space in enclosed environments. More specifically, the method finds areas of sparse sensor returns near the end of the robot’s perception in the vicinity of areas with no sensor returns and determines the directions to these areas to be the probable directions of free space due to the consistency of the lack of sensor readings with the shape of the environment. In tunnel like environments, the probable direction of free space is likely to be the center of the tunnel, while in other cases it may be the center of an open room or closer to a wall that is more heavily observed by the robot’s sensors this is because we cannot assume that merely the lack of sensor readings is the same as free space. This method can be used to assist a path planner by determining the directions of probable free space for efficient exploration.

Installation

Depends on a full installation of ROS Kinetic.

• Install and initialize ROS Kinetic desktop full, additional ROS packages, catkin-tools:

  $ sudo sh -c 'echo "deb http://packages.ros.org/ros/ubuntu `lsb_release -sc` main" > /etc/apt/sources.list.d/ros-latest.list'
  $ wget http://packages.ros.org/ros.key -O - | sudo apt-key add -
  $ sudo apt-get update
  $ sudo apt-get install ros-kinetic-desktop-full ros-kinetic-joy ros-kinetic-octomap-ros python-wstool python-catkin-tools
  $ sudo rosdep init
  $ rosdep update
  $ source /opt/ros/kinetic/setup.bash

• Initialize catkin workspace:

  $ mkdir -p ~/catkin_ws/src
  $ cd ~/catkin_ws
  $ catkin config --cmake-args -DCMAKE_BUILD_TYPE=Release
  $ catkin init  # initialize your catkin workspace

• Get the Package

  $ cd ~/catkin_ws/src
  $ git clone https://github.com/unr-arl/hfsd.git

• Build the workspace

  $ cd ~/catkin_ws
  $ catkin build 

Test Installation

• Open a terminal and type the following:

  $ cd ~/catkin_ws
  $ source devel/setup.bash
  $ roslaunch hfsd hfsd_test.launch

• In a second terminal play the test bag file included in the test directory:

  $ cd ~/catkin_ws/src/hfsd/test
  $ rosbag play test.bag

• In Rviz load the config file found in at ../hfsd/rviz_config/hfsd_test.rviz
• Rviz should show something similar to this:

Basic Usage

Launching the Package

To launch the program:

• Modify the launch file to the desired parameters.
• Run the following:

  $ cd ~/catkin_ws
  $ source devel/setup.bash
  $ roslaunch hfsd hfsd.launch

Subscribed Topics

• /cloudTransformer/inputcloud of type sensor_msgs/PointCloud2. This should be remapped to a topic that publishes point clouds. It transforms PointCloud2 messages to PCL::PointCloud<<PCL::PointXYZ>> data structures internally.
• /hfsd/odometry of type nav_msgs/Odometry. The odometry should be in the world frame. This is used to create the sliding window of PointCloud2 messages.

Published Topics

• /hfsd/OdomOut of type nav_msgs/Odometry. This simply shows the odometry being used by the algorithm.
• /hfsd/open/contours of type sensor_msgs/Image This publishes a 2D image of color coded contours extracted by the algorithm. The contours also have dots on them representing their centroids.
• /hfsd/open/image of type sensor_msgs/Image. This publishes a 2D image of the grayscale matrix that is used to determine directions of free space.
• /hfsd/visualization_marker of type visualization_msgs/MarkerArray. This publishes the visualization of the directions of free space produced by the algorithm.
• /hfsd/window_points of type sensor_msgs/PointCloud2. This publishes the sliding window of point clouds.

Paramaters

* These parameters are typically disabled and may cause unknown effects when enabled.

Contact

You can contact us for any question or remark:

• Ryan Fite
• Shehryar Khattak
• Kostas Alexis

Paper Associated With This Code

R. Fite, S. Khattak, D. Feil-Seifer and K. Alexis, "History-Aware Free Space Detection for Efficient Autonomous Exploration using Aerial Robots," 2019 IEEE Aerospace Conference, Big Sky, MT, USA, 2019, pp. 1-8. doi: 10.1109/AERO.2019.8742176

Support

This work was supported by the National Science Foundation (Award#: IIS-1757929)