projects

Robot-learning projects spanning curriculum learning, reinforcement learning, humanoid locomotion, navigation, and embodied AI systems.

robotics
project thumbnail

Curriculum & Reward Learning for Robot Navigation and Locomotion

GACL and Reward Training Wheels — adaptive training for RL robots (IROS 2025)
project thumbnail

KimoNav — from motion priors to reliable navigation

Ongoing research on execution-aware humanoid navigation, with frozen motion priors, physics experiments, and auditable results.

Scene-Aware Humanoid Locomotion and Policy Inference

Moving Through Clutter plus an ongoing ROS 2/C++ inference prototype for humanoid policies
project thumbnail

Decremental Dynamics Planning

Robot navigation that plans with full dynamics only where it matters — 2nd place in both phases of the 2025 BARN Challenge
project thumbnail

Normal Vector-Assisted SLAM Mapping

II-NVM — more accurate, more consistent maps (IEEE RA-L 2025)