Skip to content

Latest commit

 

History

History
186 lines (145 loc) · 6.63 KB

README.md

File metadata and controls

186 lines (145 loc) · 6.63 KB
Raw

Edge-SLAM

  • Authors

Ali J. Ben Ali, Zakieh Sadat Hashemifar, Karthik Dantu

  • Description

Edge-SLAM is an edge-assisted visual simultaneous localization and mapping. Edge-SLAM adapts Visual-SLAM into edge computing architecture to enable long operation of Visual-SLAM on mobile devices. This is achieved by offloading the computation-intensive modules to the edge. Thus, Edge-SLAM reduces resource usage on the mobile device and keeps it constant. Edge-SLAM is implemented on top of ORB-SLAM2.

  • Presentation Video

Presentation Video

  • Demonstration Videos

Demonstration Elevator Pitch Video

Demonstration Full Video

Publications

Ali J. Ben Ali, Zakieh Sadat Hashemifar, and Karthik Dantu. 2020. Edge-SLAM: edge-assisted visual simultaneous localization and mapping. In Proceedings of the 18th International Conference on Mobile Systems, Applications, and Services (MobiSys ’20). Association for Computing Machinery, New York, NY, USA, 325–337. DOI

License

Edge-SLAM is released under the same ORB-SLAM2 license, i.e., GPLv3 license.

For a list of all code/library dependencies (and associated licenses), please see Dependencies.md.

If you use Edge-SLAM in an academic work, please cite:

@inproceedings{10.1145/3386901.3389033,
author = {Ben Ali, Ali J. and Hashemifar, Zakieh Sadat and Dantu, Karthik},
title = {Edge-SLAM: Edge-Assisted Visual Simultaneous Localization and Mapping},
year = {2020},
isbn = {9781450379540},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3386901.3389033},
doi = {10.1145/3386901.3389033},
booktitle = {Proceedings of the 18th International Conference on Mobile Systems, Applications, and Services},
pages = {325–337},
numpages = {13},
keywords = {mobile systems, mapping, visual simultaneous localization and mapping, split architecture, localization, edge computing},
location = {Toronto, Ontario, Canada},
series = {MobiSys '20}
}

Usage

Environment Setup

  • Setup the ORB-SLAM2 prerequisites in section 2 on mobile and edge devices.

  • Setup Boost C++ Libraries on mobile and edge devices and make sure both devices have the same Boost version.

  • Our testing setup:

    • We run Edge-SLAM using the RGB-D/Stereo dataset using the ROS examples.
    • Ubuntu 18.04 LTS.
    • OpenCV 3.4.2.
    • Boost 1.65.1.
    • ROS Melodic Morenia.
    • Eigen3 3.3.4.

Building Edge-SLAM

  • Clone the repository

    git clone https://github.com/droneslab/edgeslam.git

  • Run the build script

    cd edgeslam
chmod +x build.sh
./build.sh

This will create libEdge_SLAM.so at lib folder and the executables mono_tum, mono_kitti, mono_euroc, rgbd_tum, stereo_kitti, and stereo_euroc in Examples folder.

  • To build ROS examples

    • Add the path including Examples/ROS/Edge_SLAM to the ROS_PACKAGE_PATH environment variable. Open .bashrc file and add at the end the following line. Replace PATH by the folder where you cloned Edge-SLAM

      export ROS_PACKAGE_PATH=${ROS_PACKAGE_PATH}:PATH/edgeslam/Examples/ROS

    • Run the ROS build script

      chmod +x build_ros.sh
./build_ros.sh

Running Edge-SLAM

  • Download the following TUM RGB-D dataset bag file freiburg2_desk, and adjust the DepthMapFactor in Examples/RGB-D/TUM2.yaml to 1.0.

  • On mobile device

    • Open a new terminal window, navigate to project root directory, and run

      roscore

    • Open a second terminal window, navigate to project root directory, and run

      cd Examples/ROS/Edge_SLAM/
rosrun Edge_SLAM RGBD ../../../Vocabulary/ORBvoc.txt ../../RGB-D/TUM2.yaml client

  • On edge device

    • Open a new terminal window, navigate to project root directory, and run

      roscore

    • Open a second terminal window, navigate to project root directory, and run

      cd Examples/ROS/Edge_SLAM/
rosrun Edge_SLAM RGBD ../../../Vocabulary/ORBvoc.txt ../../RGB-D/TUM2.yaml server

  • When Edge-SLAM is running on both devices, you will be asked in the Edge-SLAM terminal to enter IP and port numbers to setup the network connections. Edge-SLAM uses three TCP connections. First connection transmits keyframes, second connection transmits frames, and third connection transmits map updates. Follow the below steps in the order presented to setup the network

    • On the edge device
      • Enter the edge device IP address
      • Enter the port number to use for keyframe connection on the edge
    • On the mobile device
      • Enter the mobile device IP address
      • Enter the edge device IP address
      • Enter the port number to use for keyframe connection on the mobile
      • Enter the port number used for keyframe connection on the edge
      • If keyframe connection is established, the system will continue asking to setup the second connection
    • On the edge device
      • Enter the port number to use for frame connection on the edge
    • On the mobile device
      • Enter the port number to use for frame connection on the mobile
      • Enter the port number used for frame connection on the edge
      • If frame connection is established, the system will continue asking to setup the third connection
    • On the edge device
      • Enter the port number to use for map update connection on the edge
    • On the mobile device
      • Enter the port number to use for map update connection on the mobile
      • Enter the port number used for map update connection on the edge
      • If map update connection is established, the system on the mobile device should open the viewer windows, and the edge device should wait to receive data.

  • On the mobile device, open a third terminal window, navigate to project root directory, and run the bag file

    rosbag play ./rgbd_dataset_freiburg2_desk.bag /camera/rgb/image_color:=/camera/rgb/image_raw /camera/depth/image:=/camera/depth_registered/image_raw