VertiAdaptor: Online Kinodynamics Adaptation for Vertically Challenging Terrain
This work addresses the need for real-time terrain-aware kinodynamic adaptation in autonomous off-road navigation, offering a practical solution for improving safety and reliability in unstructured environments.
VertiAdaptor enables online adaptation of kinodynamic models for off-road autonomous driving by integrating elevation and semantic embeddings via neural ODE basis functions, improving prediction accuracy by up to 23.9% and achieving 5X faster adaptation time in simulation and on a physical platform.
Autonomous driving in off-road environments presents significant challenges due to the dynamic and unpredictable nature of unstructured terrain. Traditional kinodynamic models often struggle to generalize across diverse geometric and semantic terrain types, underscoring the need for real-time adaptation to ensure safe and reliable navigation. We propose VertiAdaptor (VA), a novel online adaptation framework that efficiently integrates elevation with semantic embeddings to enable terrain-aware kinodynamic modeling and planning via function encoders. VA learns a kinodynamic space spanned by a set of neural ordinary differential equation basis functions, capturing complex vehicle-terrain interactions across varied environments. After offline training, the proposed approach can rapidly adapt to new, unseen environments by identifying kinodynamics in the learned space through a computationally efficient least-squares calculation. We evaluate VA within the Verti-Bench simulator, built on the Chrono multi-physics engine, and validate its performance both in simulation and on a physical Verti-4-Wheeler platform. Our results demonstrate that VA improves prediction accuracy by up to 23.9% and achieves a 5X faster adaptation time, advancing the robustness and reliability of autonomous robots in complex and evolving off-road environments.