A Skeleton-Based Topological Planner for Exploration in Complex Unknown Environments

This repository contains the implementation of the paper "A Skeleton-Based Topological Planner for Exploration in Complex Unknown Environments", which is accepted by ICRA 2025.

Getting Started

The project has been tested on Ubuntu 20.04 (ROS Noetic).

sudo apt install python3-catkin-tools
mkdir -p ${YOUR_WORKSPACE_PATH}/src
cd ${YOUR_WORKSPACE_PATH}/src
git clone git@github.com:Haochen-Niu/STGPlanner.git
cd ..
catkin init
catkin build

After building, source your workspace and launch the simulation and exploration nodes:

# Launch the simulation environment
source devel/setup.zsh && roslaunch sim_env.launch  world:=Scene4

# Launch the exploration policy
source devel/setup.zsh && roslaunch exploration_manager explorer.launch world:=Scene4

# Launch the robot movement control
source devel/setup.zsh && roslaunch exploration_manager robot_move.launch

Note: To open the Gazebo GUI window during simulation, add the parameter gazebo_gui:=true to the launch commands, for example: roslaunch sim_env.launch world:=Scene4 gazebo_gui:=true

Citation

@ARTICLE{2024arXiv241213664N, author = {{Niu}, Haochen and {Ji}, Xingwu and {Zhang}, Lantao and {Wen}, Fei and {Ying}, Rendong and {Liu}, Peilin}, title = "{A Skeleton-Based Topological Planner for Exploration in Complex Unknown Environments}", journal = {arXiv e-prints}, keywords = {Computer Science - Robotics}, year = 2024, month = dec, eid = {arXiv:2412.13664}, pages = {arXiv:2412.13664}, doi = {10.48550/arXiv.2412.13664}, archivePrefix = {arXiv}, eprint = {2412.13664}, primaryClass = {cs.RO}, adsurl = {https://ui.adsabs.harvard.edu/abs/2024arXiv241213664N}, adsnote = {Provided by the SAO/NASA Astrophysics Data System} }