Recursive and Scalable 3D Coarse to Fine Path Planning
Hwajung Lee, Daegeol Ko, Jaehyuk Hur, Junwon Lee, Seongbo Ha, Jong Hwan Ko, Hyeonwoo Yu
AI summary
Problem
Path planning in large-scale, complex 3D environments faces a fundamental trade-off between path quality and computational speed, with direct 3D planning being too slow and 2D simplifications losing critical geometric information.
Approach
RUSH uses a hierarchical coarse-to-fine strategy that recursively decomposes long-range planning into independent subproblems solved in parallel by a single, unified diffusion-based network that refines initial path estimates using 3D voxel map data.
Key results
- Direct processing of large-scale 3D voxel maps without geometric information loss
- Hierarchical coarse-to-fine decomposition enabling massive parallel computation
- Unified diffusion-based network for multi-resolution path refinement
- Up to 12.59× speedup over hierarchical A* baseline with path cost within 24% of optimal
Why it matters
Provides a practical, scalable solution for rapid global navigation in complex 3D environments, benefiting autonomous vehicles, drones, and robotics applications requiring real-time replanning.
Abstract
Path planning in large-scale, complex 3D environ- ments is fundamentally constrained by a trade-off between path quality and computational speed. This paper presents RUSH (Re- cursive and Scalable 3D Coarse To Fine Path Planning), a hierar- chical framework that resolves this trade-off. RUSH decomposes the long-range planning task into a coarse plan followed by fine- grained, independent subproblems that can be solved in parallel. These subproblems are addressed by a unified, diffusion-based network that refines an initial estimate path by learning its residual to an optimal path. This approach allows RUSH to leverage rich geometric information directly from 3D voxel maps without being bottlenecked by the full map’s complexity. We validate our method on large-scale outdoor (KITTI, MulRan) and indoor (HM3D) datasets, each spanning a 200 m× 200m× 6m map. Experimental results demonstrate that RUSH generates feasible, high-quality paths with remarkable efficiency, achieving up to a 12.59× speedup over a hierarchically accelerated A∗baseline, while maintaining a path cost within 24% of the optimal solution. This performance gain positions RUSH as a powerful and practical solution for applications requiring rapid global path planning in large-scale 3D maps.