DeepGuider

DeepGuider copied to clipboard

DeepGuider

DeepGuider Project is a research project funded by national government. The DeepGuider Project aims to develop a navigation guidance system that enables robots to navigate in urban environment without the need of pre-mapping of the environment. The project was started in April 2019 and is currently under development.

DeepGuider Modules

• core: Common and basic data structures and interfaces such as pose, path, map, and algorithms
• map_manager: Map management modules to interact with cloud maps and paths
• vps: Vision algorithms to find location clues using input images and street-view images (visual positioning system; VPS)
• localizer: Localization algorithms to estimate accurate robot pose by merging various location clues
• guidance: Guidance algorithms for the robot to follow the given path and recognize its navigation states
• exploration: Exploration and recovery algorithms in degenerate cases such as lost

Dependency

• OpenCV (> 3.0.0, 3-clause BSD License)
  • OpenCV is a base for basic computer vision algorithms, linear algebra, image/video manipulation, and GUI.
• Python (> 3.6.0, PSF License)
  • Python is utilized by modules such as POI and VPS modules mainly implemented in Python language.
• QGC Ground Control Station (Apache 2.0 License)
  • UTM-related routines in QGC Ground Control Station is utilized for conversion between geodesic notation (latitude and longitude) and UTM.
  • It is included in EXTERNAL directory.

How to Run Codes

• Microsoft Windows with Microsoft Visual Studio
  • Prerequisite
    • Microsoft Visual Studio (>= 2015; for binary compatibility)
    • OpenCV binaries (v4.1.1, x64)
    • Python binaries (v3.8.5, x64)
  • Clone (or unzip) DeepGuider codes at your_folder
  • Unzip OpenCV binaries and Python binaries at your_folder
  • Open the MSVS solution file, your_folder\examples\examples.sln
  • Run any example and enjoy codes
• Linux with GCC
  • Prerequisite
    • GCC and CMake
    • OpenCV: sudo apt install libopencv-dev python3-opencv
    • Python: sudo apt install python3-dev
  • Clone (or unzip) DeepGuider codes at your_folder
  • Run ./setup_all.sh at your_folder/examples/ for initial setup (once at first after downloading)
  • Run ./build_all.sh at your_folder/examples/ for initial setup and build
  • Build an example at your_folder/examples/any_example_to_run: cmake && make
  • Run the executable and enjoy codes

DeepGuider Open Dataset

• Sidewalk video sequences with GPS and IMU tagging
• Download: https://nanum.etri.re.kr/share/roricljy/deepguider

How to Contribute Your Codes

Please refer CONTRIBUTING.md.

Authors

Please refer AUTHORS.md.

License

Please refer DeepGuider Project LSA.

Acknowledgement

The authors thank the following contributors and projects.

• This work was supported by the ICT R&D program of MSIT/IITP, Development of AI Technology for Guidance of a Mobile Robot to its Goal with Uncertain Maps in Indoor/Outdoor Environments (2019-0-01309).