NeBula-Autonomy/LAMP

Build Instructions

Install ROS

Build this package in a catkin workspace

mkdir -p catkin_ws/src
cd catkin_ws
catkin init
catkin config -DCMAKE_BUILD_TYPE=Release -DGTSAM_TANGENT_PREINTEGRATION=OFF -DGTSAM_BUILD_WITH_MARCH_NATIVE=OFF -DOPENGV_BUILD_WITH_MARCH_NATIVE=OFF -DBUILD_TEASER_FPFH=ON
cd src
git clone git@github.com:NeBula-Autonomy/LAMP.git localizer_lamp
git clone git@github.com:NeBula-Autonomy/common_nebula_slam.git
wstool init
wstool merge localizer_lamp/install/lamp_ssh.rosinstall
wstool up
catkin build lamp

The rosinstall file should take care of most of the dependencies such as GTSAM and Eigen. For the loop closure prioritization module, we also need to install some Python dependencies. This package uses python2.

torch>=1.4.0
torch-scatter==1.4.0
torch-sparse==0.4.4
torch-cluster==1.4.5
torch-spline-conv==1.1.1
torch-geometric==1.3.2
torchvision

Alternatively, run

python2 -m pip install -r requirements.txt

Note that we are using the develop branch of GTSAM, which is constantly being updated. The last tested commit of GTSAM is 99c01c4dba6443d923a28b9617b12fee06394688 for your reference.

Run Instructions

Multi-robot testing

To run a multi-robot example with our released subterranean multi-robot dataset, first download the dataset, then start the LAMP base-station process:

roslaunch lamp turn_on_lamp_base.launch robot_namespace:=base1

then play the rosbag:

rosbag play <path-to-data>/*.bag -r1 --clock clock:=/clock --wait-for-subscribers

and to visualize the map, launch rviz:

rviz -d $(rospack find lamp)/rviz/lamp_base.rviz

Data Inputs

Unit tests

To compile and run unit tests:

roscore & catkin build run_tests

To view the results of a package:

catkin_test_results build/<package_name>

Results for unit tests of packages are stored in the build/<package_name>/test_results folder.

Publications to cite when using this code

Original LAMP paper - 2020

@inproceedings{ebadi2020lamp,
  title={LAMP: Large-scale autonomous mapping and positioning for exploration of perceptually-degraded subterranean environments},
  author={Ebadi, Kamak and Chang, Yun and Palieri, Matteo and Stephens, Alex and Hatteland, Alex and Heiden, Eric and Thakur, Abhishek and Funabiki, Nobuhiro and Morrell, Benjamin and Wood, Sally and others},
  booktitle={2020 IEEE International Conference on Robotics and Automation (ICRA)},
  pages={80--86},
  year={2020},
  organization={IEEE}
}

LAMP 2.0 paper - 2022 - Overall LAMP system

@article{chang2022lamp,
  title={LAMP 2.0: A robust multi-robot SLAM system for operation in challenging large-scale underground environments},
  author={Chang, Yun and Ebadi, Kamak and Denniston, Christopher E and Ginting, Muhammad Fadhil and Rosinol, Antoni and Reinke, Andrzej and Palieri, Matteo and Shi, Jingnan and Chatterjee, Arghya and Morrell, Benjamin and others},
  journal={IEEE Robotics and Automation Letters},
  volume={7},
  number={4},
  pages={9175--9182},
  year={2022},
  publisher={IEEE}
}

Loop Closure Prioritization

@article{denniston2022loop,
  title={Loop closure prioritization for efficient and scalable multi-robot SLAM},
  author={Denniston, Christopher E and Chang, Yun and Reinke, Andrzej and Ebadi, Kamak and Sukhatme, Gaurav S and Carlone, Luca and Morrell, Benjamin and Agha-mohammadi, Ali-akbar},
  journal={IEEE Robotics and Automation Letters},
  volume={7},
  number={4},
  pages={9651--9658},
  year={2022},
  publisher={IEEE}
}

NeBula-Autonomy / LAMP