Pranay Mathur

Georgia Institute of Technology, Birla Institute of Technology and Science, Pilani - Goa Campus

Papers

Total Citations

H-Index

About

Pranay Mathur is a researcher at the forefront of robotics and embodied AI, with key contributions spanning imitation learning, human-robot interaction, and active 3D mapping. His most impactful work, **EgoMimic** (2025, 9 citations), introduces a full-stack framework that scales robot manipulation by leveraging egocentric human videos and 3D hand tracking—dramatically reducing the need for costly robot-specific demonstrations. This breakthrough addresses a core bottleneck in imitation learning, enabling robots to learn complex tasks from abundant human embodiment data. Mathur also pioneered non-invasive **Brain-Computer Interfaces (BCI) for quadcopter control** (2020, 9 citations), combining SVM and recursive least squares estimation on ROS to create intuitive human-drone interaction. In active mapping, his **Neural Visibility Field (NVF)** (2024, 4 citations) provides a principled uncertainty quantification method for Neural Radiance Fields, guiding autonomous exploration by identifying regions where the model’s predictions are unreliable due to limited visibility. With over 25 citations across his top papers, Mathur’s work bridges human data and robotic learning, pushing toward scalable, real-world deployment. His research is essential reading for anyone interested in data-efficient imitation learning, human-robot interfaces, or uncertainty-aware 3D perception.