[ICRA2025] GA-TEB: Goal-Adaptive Framework for Efficient Navigation Based on Goal Lines

This is the source code of the GA-TEB (the fourth version of the GraphicTEB series) GA-TEB: Goal-Adaptive Framework for Efficient Navigation Based on Goal Lines, an approach for robot motion planning.

Table of Contents

• Installation
• Quick Start
• Introduction for Parameters
• Contributors
• Previous Versions
• Acknowledgement

1 Installation

The project has been tested on Ubuntu 18.04 (ROS Melodic). To install the repository, please install some dependence firstly:

sudo apt install ros-melodic-navigation
sudo apt install ros-melodic-teb-local-planner

Then install OpenCV according the Chinese reference or English reference1, English reference2.

Then please install this project and build it:

mkdir -p GATEB_ws/src
cd GATEB_ws/src
git clone https://github.com/Chris-Arvin/GraphicTEB-series.git
cd ..
rosdep install --from-paths src --ignore-src --rosdistro melodic -y
[set the OpenCV_DIR in src/teb_local_planner/CMakeLists.txt according to the real location of your OpenCV]
catkin_make

2. Quick Start

See the performance of GA-TEB in preset scenes with numerous obstacles:

source GATEB_ws/devel/setup.bash
roslaunch move_base demo1_navigation.launch

Or try another demo with multiple pedestrians:

source GATEB_ws/devel/setup.bash
roslaunch move_base demo2_navigation.launch

Then, select a Nav_goal in rviz and see the performance of the robot~!

3. Introduction for Key Parameters

demo1_pedsim_simulator.launch / demo2_pedsim_simulator.launch: @param: person_mode

• 0: drive the pedestrian with data replay
• 1: drive the pedestrian with extended social force model
• 2: drive the pedestrian with manual control

@param: robot_mode

• 0: drive the robot with algorithms (baselines or your own algorithm) in the format of the plugin
• 2: drive the robot with manual control

@param: scene_file

• the localization of a .xml file describing the obstacle distribution, pedestrian distribution.

@param: pose_initial_x, pose_initial_y, pose_initial_theta

• the initial position and orientation of the robot.

4. Contributors

• Qianyi Zhang [email protected]
• Zeqing Zhu
• Zhengxi Hu
• Yinuo Song
• Yifan Yang
• Shichao Wu

5. Previous Versions

• The first version, Graphic-TEB, proposes a framework that groups obstacles with computer graphics.
• The second version, STC-TEB, adds incremental optimization to this framework.
• In parallel, another version, CG3, introduces the human gaze to make the robot keep safe distance with human beings adaptively.

6. Acknowledge

This work is based on several open-source works, thanks for their contribution and inspiration:

• navigation
• teb_local_planner
• pedsim_ros
• hdl_people_tracking