Seungchan Kim
·
Omar Alama
·
Dmytro Kurdydyk
·
John Keller
Nikhil Keetha
·
Wenshan Wang
·
Yonatan Bisk
·
Sebastian Scherer
Paper | Project Page | Video
- 01/31/2026: RAVEN was accepted to IEEE ICRA 2026!
- 10/25/2025: RAVEN was presented at IROS 2025 Active Perception Workshop, and received Oustanding Best Paper Award Finalist!
- 09/28/2025: arXiv paper is uploaded. Stay tuned for code release!
RayFronts serves as a perception and representation backbone of RAVEN.
AirStack serves as an aerial robot autonomy stack, which supports interface for Isaac Sim, and handles sensors, global & local planning, collision avoidance.
+-------------+ Interface +-------------------+ Sensor Topics +-------------------+
| Isaac Sim | -----------> | AirStack | ------------------> | RayFronts |
| Simulation | | Autonomy Stack | | Perception & Map |
+-------------+ |-------------------| |-------------------|
| - Sim interface | | - 3D Mapping |
| - Sensor drivers | | - Representation |
| - Global planner | <-------------------| - Semantic logic |
| - Local planner | Global Waypoints | - Behavior Tree |
| - Collision avoid | | |
+-------------------+ +-------------------+
First, for RayFronts setup, please follow the instructions in: RayFronts Setup
Next, for AirStack setup, please follow the instructions in:
In one terminal, start the RayFronts docker container
./run_docker.sh
Then,
./mapping_server_rosnode.sh
On another terminal, start the AirStack with IsaacSim
xhost +
airstack up