Research

My research focuses on learning-based robotic manipulation, spanning human demonstrations, multimodal observations, policy learning, simulation, and real-world execution.

Learning from demonstration

My work uses human demonstrations to learn manipulation behaviors, with diffusion policies connecting visual observations and robot state to actions.

Manipulation & touch

I study robotic interaction with irregular, deformable objects and tools, including custom grippers and tactile feedback for contact-aware control.

Simulation to real robots

My research connects simulated manipulation environments, teleoperation, and policy learning with real robotic systems.

Research systems

My work includes ChicGrasp, which combines a custom dual-jaw gripper and imitation learning for handling delicate, irregular bio-products, and deformable-object manipulation in NVIDIA Isaac Lab using teleoperated demonstrations and diffusion policies.

I have also coauthored research on tactile feedback and contact-aware control for deformable swabbing tools, and a survey of imitation learning in agricultural robotics.