Papers
7
Total Citations
272
H-Index
3
About
Blake Wulfe is a leading researcher at the intersection of robotics, machine learning, and autonomous navigation, known for pioneering large-scale, data-driven approaches to robot manipulation and motion forecasting. His work on the **Open X-Embodiment** project (119 citations) and the **DROID** dataset (108+ citations) has been instrumental in establishing the foundations for generalist robotic policies, demonstrating that training on diverse, cross-embodiment datasets can dramatically improve policy robustness and generalization. Wulfe’s research addresses critical challenges in safe, interactive autonomy, particularly through his development of **heterogeneous-agent trajectory forecasting** that incorporates class uncertainty (38 citations) and **risk-aware prediction (RAP)** for robust planning. These contributions directly tackle the long-tail safety problems inherent in real-world robot navigation and human-robot collaboration. More recently, his work on **ProVox** explores personalization and proactive planning for situated collaboration, while **GHIL-Glue** leverages generative models for hierarchical control. With a research portfolio that spans from foundational datasets to cutting-edge planning algorithms, Wulfe is shaping the next generation of capable, safe, and adaptive robotic systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2DROID: A Large-Scale In-The-Wild Robot Manipulation Dataset108 citations · 2024
- 3Heterogeneous-Agent Trajectory Forecasting Incorporating Class Uncertainty38 citations · 2022
- 4DROID: A Large-Scale In-The-Wild Robot Manipulation Dataset3 citations · 2024
- 5RAP: Risk-Aware Prediction for Robust Planning2 citations · 2022
- 6
- 7GHIL-Glue: Hierarchical Control with Filtered Subgoal Images1 citations · 2025