XIT: An Active Gas Mapping Planner for Unknown Environments

Mal Fazliu and other authors propose XIT, a sampling-based planner that models active gas distribution mapping (GDM) in unknown cluttered environments as an informative path planning problem over the next-best trajectory: a single shared tree grows simultaneously toward multiple frontier goals, performing informative gas sampling along the way while exploring unresolved regions. The authors report that high-fidelity simulations show XIT has advantages in GDM quality and efficiency, and that scaled-down real-world experiments show its practicality. The authors state that although the method was developed for active GDM, it is also applicable to other robotic information-gathering tasks in unknown environments that face an exploration-exploitation trade-off.

AI 根据报道生成 · 2 days ago更新