Real-Time Support Terrain Mapping and Terrain Adaptive Local Planning for Quadruped Robots

The expanding application of quadruped robots in complex environments poses new challenges to their perception and planning systems. This letter proposes a novel approach for local perception and planning, which integrates exteroception, such as laser point clouds and depth images, with proprioception such as foot force and achieves terrain elevation mapping with dense support surface prediction and quadruped terrain traversability analysis. Furthermore, it enhances traditional 2D state lattice planning with terrain interaction prediction, enabling perception-based decision-making at the high-level command adaptive adjustment. The proposed system was tested in both indoor and outdoor environments, and achieved good performances in support surface prediction and adaptive local planning and locomotion.