Papers, awards, people, and talks from the lab. Newest first.
The PhD student gave a Rising Star talk at the Modeling, Estimation and Control Conference in Arizona, and presented MAGMA in a regular session.
Adds constraint handling to heat-flow trajectory optimization in a way that provably stays a gradient flow, so it needs far less tuning.
Plans contact-rich motions for legged robots and humanoids with up to 32 joints, and was the only method to solve every benchmark problem. Shown on a Unitree Go2 playing hopscotch.
AirSplan, TrackDeform3D, and Phasing Through the Flames. AirSplan is undergraduate William Hong's first paper.
The PhD student received the College of Engineering's J. Robert Beyster Computational Innovation Graduate Fellowship.
Trains a vision-language-action model to navigate long routes as a multi-turn conversation, and runs zero-shot on a real robot. Led by postdoc Yue Hu.
Recognized for research in legged robotics and control, and for helping build the department's graduate programs.
An interior point method for trajectory optimization of rigid bodies on Lie groups, led by alum Sangli Teng, now at UC Berkeley.
Computes how uncertain a NeRF or Gaussian splat is about color, depth, and labels with no extra training, so a robot can decide where to look next.
The NeurIPS 2025 paper on tracking a robot's state through nonlinear dynamics and non-Gaussian noise is a finalist for the IEEE RAS Technical Committee on Optimization for Robotics 2025 Best Paper Award.
Closes the loop between human mesh recovery and Gaussian splatting avatars. Led by Yeheng Zong with Seth Isaacson and Yizhou Chen.
Real-time 3D semantic mapping that uses much less memory and handles fixed and open-vocabulary labels in one framework.
Predicts how branched cables and wire harnesses bend and move in real time, so a robot can track and manipulate them.
Maps shipwrecks in real time from an underwater robot's stereo cameras, and releases over 24,000 underwater stereo pairs. Conditionally accepted to IEEE Transactions on Field Robotics.
He is one of the inaugural members of the center's steering committee.
Challen defended his PhD on heat-flow trajectory optimization and is now a research engineer at Amazon Robotics.
Lets a robot arm work out the mass and inertia of an unknown payload by moving it, while guaranteeing it stays within its limits and clear of obstacles. Presented by Bohao Zhang.
Both defended their PhDs in March. Bohao is now at Tesla Optimus, and Zachary is a postdoc at Yale.
Xun Fu and Bohao Zhang, with Talia Moore's lab, found the tail design that best turns a body in midair. In the Journal of the Royal Society Interface.
Older news lives in the publications list and on Ram's Michigan Robotics profile.
