Botao He
Botao He is a PhD student in Computer Science at the University of Maryland, College Park, where his research focuses on interactive perception and learning, dexterous manipulation, and event-camera-based robotics. He is best known for his work on the FEEL (Force-Enhanced Egocentric Learning) dataset, which captures surface electromyography signals from a wearable wristband to estimate per-finger forces for human demonstration learning. He has also published on event-based 3D Gaussian splatting for robot egomotion, aerial navigation, and active human pose estimation via autonomous UAV agents. He has held a research internship at Amazon FAR and was previously a research assistant at the FAST Lab at Zhejiang University.
“The paper trains a 'play policy' from minutes of goal-free human teleoperation, then lets the robot run autonomously for hours collecting contact-rich data.”
Source→“The core contribution is a $300 wristband that captures surface electromyography (sEMG) signals from the forearm and converts them into per-finger force estimates, enabling force-enriched human demonstrations without instrumenting the fingertips.”
Source→AI-extracted from podcast / newsletter / paper summaries. May contain errors.