High-speed robot navigation improves with self-supervised traversability mapping
TravSUITE: Traversability via Self-Supervised, Uncertainty-Aware IRL and Terrain Estimation
Robotics
Summary
Navigating rough outdoor terrain is tough for robots because they need to understand the environment and know where it’s safe to drive. The authors of this paper created TravSUITE, a system that helps robots make detailed maps using cameras and sensors without needing manual labels. Their approach predicts which paths the robot can take safely by learning from the data itself and estimating uncertainty to avoid risky areas. Tests show that this method leads to fewer navigation errors and faster travel, especially in challenging, off-road conditions.
What this means in practice
- •For robotics development teams: Improve autonomous off-road vehicle navigation by integrating a self-supervised system that builds detailed local maps and predicts safe paths with uncertainty awareness.
- •For drone operators: Use uncertainty-aware terrain mapping to enhance drone route planning in complex natural environments with limited sensor information.
Authors
Samuel Triest, Amirreza Shaban, David D. Fan, Wenshan Wang, Sebastian Scherer
Abstract
Traversability analysis in off-road settings remains a fundamental challenge for mobile robots. Key difficulties include constructing an accurate and expressive local map from multimodal sensor data, and using the map to design traversability rules that yield desirable navigation behavior. Importantly, this system must be resilient to the limited sensing regime brought about by complex environments and high speeds. In this paper, we present TravSUITE, a traversability system suitable for high-speed navigation in off-road environments. TravSUITE consists of two major components: 1) an efficient voxel mapper that leverages visual foundation models (VFMs) to build a rich geometric-semantic local map from streams of on-board sensor data, and 2) a unified neural network that jointly predicts traversability-relevant quantities in bird's eye view (BEV), including geometry, semantics, speed and cost. Our training strategy is entirely annotation-free and self-supervised, leveraging tasks such as map inpainting and inverse reinforcement learning (IRL) to learn both map representations and traversability. We perform a thorough ablation study and comparison to state-of-the-art approaches, and the results indicate that cost learning and auxiliary inpainting each contribute significantly to planning quality, and their combination is critical for achieving state-of-the-art performance in off-road path planning. We also design a simple risk adaptation mechanism to leverage our method's uncertainty estimates at deploy-time, and demonstrate that a combination of inpainting and risk estimation can result in 80% fewer navigation errors and 5% faster autonomous traversal speeds in real-world hardware experiments.