ScanDreamer: A Map-Grounded World Model for Quadrotor Navigation
ScanDreamer: A Map-Grounded World Model for Quadrotor Navigation
This paper introduces ScanDreamer, a world model designed for quadrotor navigation in cluttered, unknown environments.
The study presents ScanDreamer, a map-grounded world model for quadrotor navigation that combines explicit geometric memory with perception-aware motion constraints to achieve 96.3% success and 3.6% collision rates in unseen environments within NVIDIA Isaac Sim.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
episodes
>1,560 episodes
Reported trials
View original evidence
tes the quadrotor pose over a frozen map and derives progress and wall-contact signals from the mapped geometry. We train ScanDreamer for 0.8 million environment steps (5.5 h on one GPU) and evaluate it in NVIDIA Isaac Sim over 1560 episodes spanning unseen layouts and parameter ranges. Across three seeds (mean $\pm$ SD), ScanDreamer achieves 96.3 $\pm$ 0.3% success with 3.6 $\pm$ 0.3% collision, compared with 59.6 $
Open source S4
- code: Available Artifact link Open source S5
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
tes the quadrotor pose over a frozen map and derives progress and wall-contact signals from the mapped geometry. We train ScanDreamer for 0.8 million environment steps (5.5 h on one GPU) and evaluate it in NVIDIA Isaac Sim over 1560 episodes spanning unseen layouts and parameter ranges. Across three seeds (mean $\pm$ SD), ScanDreamer achieves 96.3 $\pm$ 0.3% success with 3.6 $\pm$ 0.3% collision, compared with 59.6 $
Open source S4
arXiv:2611.11605v1 Announce Type: new Abstract: Quadrotors inspecting buildings and tunnels must navigate unknown clutter with a forward-facing depth camera whose limited field of view constrains scene understanding and collision-avoiding navigation. Recent world-model agents learn navigation through imagined experience (e.g., AirDreamer), but remain data-intensive and often lack explicit geometric memory and awarene
Open source S2
\pm$ 3.0% success and 28.8 $\pm$ 2.8% collision for AirDreamer at the same online-interaction budget. Removing the forward-cone penalty increases collisions to 9.9 $\pm$ 3.1%. These results show that combining explicit map memory, map-grounded imagination, and perception-aware motion substantially improves navigation performance and data efficiency. The source code is available at https://github.com/WANGSHUAIzzzz/Sca
Open source S5
Source：arXiv Robotics — research abstracts · arxiv.org