TRAVEL for RA-L'22 w/ IROS Option

Best Paper Award winner from RA-L 2022

Official page of "TRAVEL: Traversable Ground and Above-Ground Object Segmentation using Graph Representation for 3D LiDAR Scans", which is accepted by RA-L with IROS'22 option.

Demo

Keywords

Object segmentation, Traversable ground segmentation, Graph search, Autonomous navigation, LiDAR

To Do

This archive is still being updated. Up until the IROS'22 presentation, I will update the final version.

runnable node on general setting.
Add the application codes
Update for docker users

Contributors

Minho Oh & Euigon Jung
email:
- Minho Oh: minho.oh@kaist.ac.kr
- Euigon Jung: euigon94@gmail.com
Advised from Hyungtae Lim and Hyun Myung @ KAIST

Test Env.

Ubuntu 18.04 LTS
ROS Melodic

How to Build

Dependencies

sudo apt install cmake libeigen3-dev libboost-all-dev
sudo apt-get install ros-melodic-jsk-recognition
sudo apt-get install ros-melodic-jsk-common-msgs
sudo apt-get install ros-melodic-jsk-rviz-plugins

Build

mkdir -p catkin_ws/src/
cd catkin_ws/src/
git clone https://github.com/url-kaist/TRAVEL.git
../
catkin_make

How to Run In KITTI

roslaunch travel run_kitti.launch data_dir:=${ABS_DIRECTORY_OF_KITTI} seq:=${SEQUENCE_ID}
# for example,
# then, the bin files in `/home/shapelim/qbench/kitti/00/velodyne` are loaded in the ascending order 
roslaunch travel run_kitti.launch data_dir:=/home/shapelim/qbench/kitti seq:=00

Download our dataset

rough_terrain_1 (24.5GB) and rough_terrain_2 (15.4GB) are our rough terrain dataset.

wget https://urserver.kaist.ac.kr/publicdata/TRAVEL_22/rough_terrain_1.bag
wget https://urserver.kaist.ac.kr/publicdata/TRAVEL_22/rough_terrain_2.bag

How to Run TRAVEL

RUN!
```
roslaunch travel run.launch
```

On your setting.

Include two header files in your source. "tgs.hpp" & "aos.hpp"
Initialize "travel::TravelGroundSeg" and "travel::ObjectCluster"
Use the "setParams()" function in each class to set the parameters.
Use "travel::TravelGroundSeg.estimateGround()" function for traversable ground segmentation
Use "travel::ObjectCluster.segmentObjects()" function for above-ground object segmentation

I will upload an example ros node that subscribes to sensor data.
If you want to use TRAVEL with python code, then visit here (https://github.com/darrenjkt/TRAVEL). Thank you Darren :)

Aplications

Local terrain map.
Global traversable map.
Global static traversable map by using TRAVEL with ERASOR

Citation

If our research has been helpful, please cite the below papers:

@ARTICLE{oh2022travel,  
    author={Oh, Minho and Jung, Euigon and Lim, Hyungtae and Song, Wonho and Hu, Sumin and Lee, Eungchang Mason and Park, Junghee and Kim, Jaekyung and Lee, Jangwoo and Myung, Hyun},  
    journal={IEEE Robotics and Automation Letters},   
    title={TRAVEL: Traversable Ground and Above-Ground Object Segmentation Using Graph Representation of 3D LiDAR Scans},   
    volume={7},  
    number={3},  
    pages={7255-7262},  
    year={2022},
    }

@article{lim2021patchwork,
    title={Patchwork: Concentric Zone-based Region-wise Ground Segmentation with Ground Likelihood Estimation Using a 3D LiDAR Sensor},
    author={Lim, Hyungtae and Minho, Oh and Myung, Hyun},
    journal={IEEE Robot. Autom. Lett.},
    volume={6},
    number={4},
    pages={6458--6465},
    year={2021},
    }

@article{lim2021erasor,
    title={ERASOR: Egocentric Ratio of Pseudo Occupancy-Based Dynamic Object Removal for Static 3D Point Cloud Map Building},
    author={Lim, Hyungtae and Hwang, Sungwon and Myung, Hyun},
    journal={IEEE Robotics and Automation Letters},
    volume={6},
    number={2},
    pages={2272--2279},
    year={2021},
    publisher={IEEE}
    }

url-kaist / TRAVEL