Papers
14
Total Citations
401
H-Index
9
About
Anushri Dixit is a robotics researcher whose work sits at the intersection of autonomous navigation, risk-aware planning, and embodied AI. She is best known for her contributions to off-road robotic autonomy, particularly through the development of STEP (Stochastic Traversability Evaluation and Planning), a framework enabling robots to assess and navigate hazardous, unstructured terrains with probabilistic risk awareness — work that has garnered over 110 citations and proven influential in real-world deployment contexts. Dixit was also a key contributor to TEAM CoSTAR's NeBula system, the award-winning robotic autonomy stack that secured victory in Phase II of the prestigious DARPA Subterranean Challenge, with related publications collectively accumulating over 150 citations. Her research extends to legged robot stability through predictive proprioception and failure detection in extreme environments. More recently, she has pioneered work at the frontier of large language models in robotics, developing KnowNo — a framework that equips LLM-based planners with calibrated uncertainty to know when to ask for help — alongside exploration strategies for embodied question answering. Across her portfolio, Dixit consistently addresses the challenge of making robots reliably intelligent in the messiness of the real world.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10