YI LIU
Yi Liu, M.S @ UMich

Ford Motor Robotics Building
University of Michigan
Ann Arbor, MI, 48105
I am second year master student at the University of Michigan Electrical & Computer Engineering Department, specializing in Robotics . I am closely working with Prof.Katie Skinner and Jingyu Song at the Field Robotics Group and Ford Center for Autonomous Vehicle. Prior to that, I received my Bachelor of Engineering degree from Glasgow College, a joint institute of the University of Electronic Science and Technology of China (UESTC) and the University of Glasgow.
My research focuses on integrating cutting-edge machine learning and computer vision to solve fundamental challenges in robotic perception and navigation. My primary interests include:
-
Robust Multi-Modal Perception: I develop robust perception systems that fuse data from multiple sensors—such as cameras, LiDAR, and radar—to ensure safe and efficient navigation in dynamic and unstructured environments. The goal is to create systems that are resilient to individual sensor failures and adverse conditions.
-
Foundation Models for Robotic Navigation: I investigate how large-scale vision and vision-language foundation models can be adapted for robotics navigation. By leveraging the rich world knowledge embedded in these foundation models, my work aims to equip robots with a deep, semantic understanding of their surroundings while dramatically reducing the need for costly, hand-labeled training data.
news
Nov 11, 2024 | Excited to join the Field Robotics Group as a Research Assistant! 🐸 |
---|---|
Aug 26, 2024 | Started my M.S. in Robotics at the University of Michigan — Go Blue! 💙〽️ |
Jun 01, 2024 | I graduated from Glasgow College, a joint institute of the University of Electronic Science and Technology of China (UESTC) and the University of Glasgow, with a First Class Honours Degree! 🥳 |
selected publications
- FishDetector-R1FishDetector-R1: Unified MLLM-Based Framework with Reinforcement Fine-Tuning for Weakly Supervised Fish Detection, Segmentation, and Counting2026Submitted to WACV 2026