Yuchen Zhang

I am a Ph.D. student in Robotics at Carnegie Mellon University, advised by Dr. Wenshan Wang and Dr. Sebastian Scherer. My research focuses on building generalizable and efficient perception systems for robotics. Previously, I worked on correspondence prediction, visual odometry, and reinforcement learning for quadrupedal locomotion. I earned my M.S. in Robotics at CMU and my B.S. in Electrical Engineering at UC San Diego. I also spent 7 years competing in the FIRST Robotics Competition (FRC) and FIRST Tech Challenge (FTC).

Email  /  CV  /  Scholar  /  Twitter  /  Github

profile photo

Research

I am interested in perception and hardware integration for general-purpose robots. Most of my papers focus on perception foundation models. Some papers are highlighted.

MapAnything: Universal Feed-Forward Metric 3D Reconstruction
Nikhil Keetha, Norman Müller, Johannes Schönberger, Lorenzo Porzi, Yuchen Zhang, Tobias Fischer, Arno Knapitsch, Duncan Zauss, Ethan Weber, Nelson Antunes, Jonathon Luiten, Manuel López Antequera, Samuel Rota Bulò, Christian Richardt, Deva Ramanan, Sebastian Scherer, Peter Kontschieder
In Submission
project page / arXiv / code / model / demo

A feedforward geometric model that can utilize prior camera poses, intrinsics, and (sparse) depth maps to produce metrically accurate 3D reconstructions.

UFM: A Simple Path towards Unified Dense Correspondence with Flow
Yuchen Zhang, Nikhil Keetha, Chenwei Lyu, Bhuvan Jhamb, Yutian Chen, Yuheng Qiu, Jay Karhade, Shreyas Jha, Yaoyu Hu, Deva Ramanan, Sebastian Scherer, Wenshan Wang,
NeurIPS, 2025
project page / arXiv / code / model / demo

A simple, transformer model that can benefit from combining optical flow and wide-baseline matching data, out-performing specialized models on both domain.

Vision-Guided Quadrupedal Locomotion in the Wild with Multi-Modal Delay Randomization
Chieko Sarah Imai*, Minghao Zhang*, Yuchen Zhang*, Marcin Kierebiński, Ruihan Yang, Yuzhe Qin, Xiaolong Wang,
IROS, 2022
project page / arXiv / code

A domain randomization strategy that simulates multi-modal delays in real hardwares, enabling sim2real transfer of locomotion policies.

Projects
FOC Motor Controller

A single channel Field-Oriented Control (FOC) motor controller based on STM32. Achieved 2kHz current loop bandwidth.

Hand-held Gimbal

A 3-channel FOC controller that stabilizes the phone based on IMU readings.


This website is based on template at here.