SE(2) NavMesh helps robots navigate tight spaces with yaw-aware planning

Researchers from ETH Zurich and other institutions have developed the SE(2) Navigation Mesh (SE(2) NavMesh), a novel approach to global navigation for ground robots operating in complex, multi-level environments. Current methods like point clouds and volumetric occupancy maps lack explicit surface structure for traversability estimation, while dense triangle meshes are computationally prohibitive for pathfinding. Traditional navigation meshes assume yaw-invariant traversability, making them unsuitable for non-circular robots in tight spaces such as corridors or doorways.

The SE(2) NavMesh addresses these limitations by encoding yaw-dependent traversability using footprint masks. It constructs a graph over yaw-specific layers with explicit translational and rotational connectivity. The team developed a hierarchical A*-String Pulling-A* (ASA) pathfinding strategy that optimizes both robot position and heading. In simulation, the SE(2) NavMesh captured over 50% more traversable area than classical NavMeshes. Extensive real-world experiments on a physical robot confirmed real-time online generation and successful navigation across multiple environments, consistently outperforming sampling-based baselines in constrained settings.