Introducing ugv_nav4d: Advanced Multi-Surface Navigation for UGVs with 4D Path Planning (ROS 2 Ready!)
ugv_nav4d is an open-source global path planner for terrestrial robots to navigate complex indoor and outdoor environments, now with a fully-supported ROS 2 wrapper.
Key highlights
- Uses a traversability map based on a Multi-Layered Surface Map (MLSMap), enabling planning in multi-surface and multi-storey environments.
- Builds paths using motion primitives tailored to the robot’s mechanical features, ensuring compatibility with its motion capabilities.
- Incorporates detailed vertical information for smoother and more accurate terrain modeling compared to conventional voxel-based approaches.
Why ugv_nav4d?
Conventional planners like the ROS 2 Navigation Stack often lose detail and accuracy in 3D and multi-level environments due to voxelization and separate costmaps.
ugv_nav4d enhances environmental representation through:
- MLS maps and detailed 3D traversability maps,
- A single TraversabilityMap3D for accurate ground surface data,
- Smoother, more realistic terrain modeling—crucial for navigation and decision-making in complex environments.
Core Features
- ARA-based Planning:* Real-time search for shortest paths using the SBPL (Search-Based Planning Library), incorporating robot and terrain-specific costs such as robot dimensions, collision, terrain steepness, and motion primitives.
- Pre-Computed Motions: Primitives for forward, backward, lateral, and point-turn movements, supporting Ackermann, differential, and omni-directional robots.
- Terrain-aware Navigation: Utilizes traversability and slope data to avoid non-traversable or risky areas, enhancing safety and performance on rough or multi-surface terrain.
- Continuous 3D Mapping: Leverages continuous multi-layer surface maps (MLSMap) for accurate modeling of complex and multi-storey environments.
- Modular and Extensible: Easily adapt or extend the core planner for new robot types, motion models, or terrain representations.
- ROS 2 Integration: Seamlessly deployable in modern ROS 2 navigation stacks via a dedicated wrapper, supporting simulation and real-robot use.
- Debugging GUI: An included PlannerGui for testing, visualization, and tuning on static maps.
Proven in the Field
ugv_nav4d has been field-tested in major research projects (ENTERN, ANT, VIPE, KIMMI-SF, HiSE, PerSim, CoRobX) and successfully deployed on a range of terrestrial robots, including Artemis, Coyote-III, Crex, Charlie, Hunter-SE, SherpaTT, and Asguard-IV.
Read More
For a full technical overview, see our peer-reviewed paper:
Böckmann et al., “ugv_nav4d: Advanced Multi-Surface Navigation for Unmanned Ground Vehicles Using 4D Path Planning Techniques”
Journal of Open Source Software, 2024, 9(104), 6983
Get Started
- Source & Docs: github.com/dfki-ric/ugv_nav4d
- ROS 2 wrapper: github.com/dfki-ric/ugv_nav4d_ros2
*Developed at the DFKI Robotics Innovation Center and University of Bremen.