Show EOL distros:
Package Summary
The spatio-temporal 3D obstacle costmap package
- Maintainer status: maintained
- Maintainer: Steve Macenski <stevenmacenski AT gmail DOT com>
- Author:
- License: LGPL v2.1
- Source: git https://github.com/SteveMacenski/spatio_temporal_voxel_layer.git (branch: kinetic-devel)
Package Summary
The spatio-temporal 3D obstacle costmap package
- Maintainer status: maintained
- Maintainer: Steve Macenski <stevenmacenski AT gmail DOT com>
- Author:
- License: LGPL v2.1
- Source: git https://github.com/SteveMacenski/spatio_temporal_voxel_layer.git (branch: melodic-devel)
Package Summary
The spatio-temporal 3D obstacle costmap package
- Maintainer status: maintained
- Maintainer: Steve Macenski <stevenmacenski AT gmail DOT com>
- Author:
- License: LGPL v2.1
- Source: git https://github.com/SteveMacenski/spatio_temporal_voxel_layer.git (branch: noetic-devel)
Contents
Overview
The spatio-temporal voxel layer incorporates information from the sensors in the form of PointClouds or LaserScans. This information is converted into 3D and populated into an efficient voxel grid for each sensor cycle.
ROS drop in replacement to the voxel layer which uses OpenVDB's lightening fast voxel representation originally created for Dreamworks' animations. It runs at significantly less CPU than the existing voxel_layer plugin using the voxel_grid structure.
Created in response to need for a more efficient voxel grid to process 10+ high resolution depth sensors in parallel and reduce the CPU load of the move_base process.
It also introduces the concept of global and local voxel decay, which temporally and spatially decays voxels rather than raycasting. We find major improvements in performance and usability in highly dynamic spaces. OpenVDB's efficient iterators and speed allow for this new technique and is suitable for outdoor environments as well as indoor. It also allows for more complex inputs to costmap_2d in the form of clustered or tracked depth blobs.
More information, ROS API, demos, and resources are given in the GitHub page.
You can find this work https://journals.sagepub.com/doi/10.1177/1729881420910530.
@article{doi:10.1177/1729881420910530,
- author = {Steve Macenski and David Tsai and Max Feinberg}, title ={Spatio-temporal voxel layer: A view on robot perception for the dynamic world}, journal = {International Journal of Advanced Robotic Systems}, volume = {17}, number = {2}, year = {2020}, doi = {10.1177/1729881420910530},
}
