ProTerrain: Probabilistic Physics-Informed Rough Terrain World Modeling

TL;DR

ProTerrain introduces a probabilistic world modeling framework that captures spatially correlated terrain uncertainties and propagates them through physics-based predictions, enhancing off-road navigation safety.

Contribution

It presents a novel probabilistic modeling approach that explicitly accounts for spatial correlations in terrain uncertainty, improving prediction reliability in unstructured environments.

Findings

Enhanced uncertainty estimation accuracy

Improved trajectory prediction performance

Efficient high-resolution probabilistic modeling

Abstract

Uncertainty-aware robot motion prediction is crucial for downstream traversability estimation and safe autonomous navigation in unstructured, off-road environments, where terrain is heterogeneous and perceptual uncertainty is high. Most existing methods assume deterministic or spatially independent terrain uncertainties, ignoring the inherent local correlations of 3D spatial data and often producing unreliable predictions. In this work, we introduce an efficient probabilistic framework that explicitly models spatially correlated aleatoric uncertainty over terrain parameters as a probabilistic world model and propagates this uncertainty through a differentiable physics engine for probabilistic trajectory forecasting. By leveraging structured convolutional operators, our approach provides high-resolution multivariate predictions at manageable computational cost. Experimental evaluation on a publicly available dataset shows significantly improved uncertainty estimation and trajectory prediction accuracy over aleatoric uncertainty estimation baselines.