• 设为首页
  • 点击收藏
  • 手机版
    手机扫一扫访问
    迪恩网络手机版
  • 关注官方公众号
    微信扫一扫关注
    迪恩网络公众号

melhousni/ndt_mapping_localization: C++ Implementation of the NDT mapping and lo ...

原作者: [db:作者] 来自: 网络 收藏 邀请

开源软件名称(OpenSource Name):

melhousni/ndt_mapping_localization

开源软件地址(OpenSource Url):

https://github.com/melhousni/ndt_mapping_localization

开源编程语言(OpenSource Language):

C++ 66.7%

开源软件介绍(OpenSource Introduction):

ndt_mapping_localization

C++ Implementation of the NDT mapping and localization algorithm for ADV on ROS.

Two packages available in this implementation :

  • vehicle_mapping : Pointcloud registration using the 3D NDT algorithm assisted by an EKF.
  • vehicle_localization : 6-DoF Localization using the 3D NDT algorithm assisted by an EKF.

NDT Results

Dependecies :

Compile :

cd ~/catkin_ws
git clone https://github.com/melhousni/ndt_mapping_localization
cd ..
catkin_make

How to use :

  • vehicle_mapping :

    1. Set up the topic names for GPS and IMU data in the vehicle_state/launch/ekf.launch and vehicle_state/param.ekf.yaml. For more information of how to use the robot_localization package, please see : http://docs.ros.org/lunar/api/robot_localization/html/index.html
    2. Set up the topic names for Pointcloud and EKF Odom data in vehicle_mapping/launch/ndt_mapping.launch
    3. Set up the path to save the final pointcloud in vehicle_mapping/launch/ndt_mapping.launch
    4. Use roslaunch vehicle_mapping ndt_mapping.launch to launch the mapping package and play the rosbag that you want to use as input in an other terminal.
  • vehicle_localization :

    1. Set up the topic names for Pointcloud and EKF Odom data in vehicle_localization/launch/map_localization.launch
    2. Set up the path to load the pointcloud map in vehicle_localization/launch/map_localization.launch
    3. Choose which method to use to find the initial position. (RVIZ or GPS)
      If you choose RVIZ, you will have to give a guess to the localization package on where to start using the "2D Pose Estimate" button in RVIZ
      If you choose GPS, you will have to give the coordinates of the origin of your pointcloud map and set them up in vehicle_localization/launch/map_localization.launch.
      At the moment, the method based on RVIZ is more stable than the GPS one
    4. Launch a rosbag throught another terminal, or launch live data streaming from the sensors.

Demo

We provide a pointcloud map built using the mapping package, and a rosbag recorded on a different day to test the localization package, which you can download from here : https://drive.google.com/open?id=1mqBURRoouFT1HYkWjkjENcotXf75Tz0u

If you still have problem running the packages, you can check the following videos demonstration how to use both the RVIZ and GPS approchaes :

Notes

  • Sometimes, slowing down the rosbag playing speed (using -r) can improve the results of the localization package
  • Tweaking the scoreThreshold parameter can also improve results

Please do not hesitate to report any problems or bugs you may encounter, either throught the issues page of this repo, or by sending me an email at melhousni@wpi.edu.

Citing

If you use theses packages, please consider refer to this repo with :

@misc{housni2019,
author = {Mahdi, Elhousni},
title = {NDT-EKF Mapping & Localization},
year = {2019},
publisher = {GitHub},
journal = {GitHub repository},
howpublished = { https://github.com/melhousni/ndt_mapping_localization }
}




鲜花

握手

雷人

路过

鸡蛋
该文章已有0人参与评论

请发表评论

全部评论

专题导读
热门推荐
阅读排行榜

扫描微信二维码

查看手机版网站

随时了解更新最新资讯

139-2527-9053

在线客服(服务时间 9:00~18:00)

在线QQ客服
地址:深圳市南山区西丽大学城创智工业园
电邮:jeky_zhao#qq.com
移动电话:139-2527-9053

Powered by 互联科技 X3.4© 2001-2213 极客世界.|Sitemap