Samuel Triest
Papers
5
Total Citations
115
H-Index
4
About
Samuel Triest is a robotics researcher whose work centers on autonomous navigation in off-road and unstructured environments, with particular expertise in traversability estimation, costmap learning, and self-supervised learning for field robotics. His most influential contribution, "How Does It Feel? Self-Supervised Costmap Learning for Off-Road Vehicle Traversability" (2023, 56 citations), introduced a novel approach to learning terrain traversability without the burden of hand-crafted labels, leveraging the physical interaction between robot and terrain as a natural supervisory signal. This work exemplifies his broader commitment to reducing the engineering overhead traditionally required for off-road autonomy. Triest further expanded the field's empirical foundations through TartanDrive 2.0 (2024, 25 citations), a large-scale multimodal dataset that has become a valuable resource for the off-road driving research community. His work on risk-aware costmaps via inverse reinforcement learning (2023, 23 citations) demonstrates his ability to integrate principled uncertainty reasoning into practical navigation systems. Across his portfolio, Triest consistently bridges the gap between theoretical machine learning advances and the harsh realities of real-world field robotics, establishing himself as a promising voice in next-generation autonomous vehicle research.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4Rough Terrain Navigation Using Divergence Constrained Model-Based Reinforcement Learning7 citations · 2022
- 5