Awesome

hdl_global_localization

on ROS melodic and noetic

Requirements

hdl_global_localization requires the following libraries:

• PCL
• OpenCV
• OpenMP
• Teaser++ [optional]

Services

• /hdl_global_localization/set_engine (hdl_global_localization::SetGlobalLocalizationEngine)
  • Available global localization engines: BBS, FPFH_RANSAC, FPFH_TEASER
• /hdl_global_localization/set_global_map (hdl_global_localization::SetGlobalMap)
• /hdl_global_localization/query (hdl_global_localization::QueryGlobalLocalization)

Algorithms

• 2D Grid Map-based Branch-and-Bound Search
  • Real-time loop closure in 2D LIDAR SLAM, ICRA, 2016
• FPFH + RANSAC (based on pcl::SampleConsensusPrerejective)
  • Fast Point Feature Histograms (FPFH) for 3D registration, ICRA, 2009
  • Pose Estimation using Local Structure-Specific Shape and Appearance Context, ICRA, 2013
• FPFH + Teaser++
  • TEASER: Fast and Certifiable Point Cloud Registration, T-RO, 2020

Related packages

• interactive_slam
• hdl_graph_slam
• hdl_localization
• hdl_people_tracking

Contact

Kenji Koide, k.koide@aist.go.jp

Human-Centered Mobility Research Center, National Institute of Advanced Industrial Science and Technology (AIST), Japan