Past Projects

I've worked on several AI and robotics projects in the past. Below I've outlined some focus areas for more recent work (things in the past ~3 years) with links to the relevant projects. This page doesn't include work I've done in industry.

Task and Motion Planning for Mobile Manipulation

With the Progress Lab at Michigan, I worked on approaches to do task and motion planning for mobile manipulation with the Digit humanoid robot.

Optimal Constrained Task Planning as Mixed Integer Programming
Published in IROS 2022
Probabilistic Inference in Planning for Partially Observable Long Horizon Problems
Published in IROS 2021

Legged Locomotion

Most recently, I spent a year working on legged locomotion for the Digit bipedal humanoid robot with the ROAHM Lab at Michigan. I can't say too much about these yet since they're still in the process of being published, but I'll update their pages when they're out.

Provably Safe Motion Planning for Humanoid Bipedal Locomotion
In progress.
Reinforcement Learning for Humanoid Bipedal Locomotion
In progress.

Natural Language for Embodied AI

I've worked on a few projects related to embodied AI with the SLED Group at Michigan. These projects were focused on the ability of agents in home environments to follow natural language instructions in conjunction with perception and planning.

Deliberative Agent for Following Natural Language Instructions
Published in EMNLP 2022
SEAGULL: An Embodied Agent for Instruction Following Through Situated Dialog
2023 Amazon Alexa Prize Simbot Challenge

Representation Learning for Robotics

I've done some work on representation learning for robotics as well.

Topologically-Informed Atlas Learning
Published in ICRA 2022