| Yuyang Gao | Academic Homepage |
Yuyang Gao高瑀阳
Incoming M.S. Student · School of Computer Science · Peking University
I am interested in robotic systems that can live and operate in the real world for extended periods — not merely executing isolated instructions, but continuously interacting with their environments, accumulating experience from physical feedback, and improving their behavior over time. My current work uses wet-lab experimentation as a concrete testbed for long-horizon, self-improving embodied agents.
🎓Education
Peking University, incoming M.S. student in Computer Science; research on robotic perception, manipulation, and closed-loop experimental systems.
Dalian University of Technology, B.Eng. student in Automation; GPA 94.6; ranked 1st out of 162; national scholarship recipient.
🧪Research
Embodied agents. Building robots that perceive, plan, execute, and reflect in real environments.
Closed-loop science. Using physical feedback such as mass, pH, and conductivity to support adaptive experimental workflows.
Robot manipulation. Skill abstraction, precise pouring, grasping, and long-horizon execution.
Multimodal perception. Connecting VLMs, pointing models, segmentation, and depth-based pose solving.
💼Experience
Peking University · School of Computer Science
Research intern on robotic perception, manipulation, and closed-loop experimental systems.
Tsinghua University · AI Industry Research Institute
Research intern on dual-arm manipulation and imitation learning with simulated and real robot demonstrations.
Dalian University of Technology · Automation
Undergraduate training in automation, control, robotics, and intelligent systems.
🧩Selected Projects
A wet-lab robotic platform that connects perception, planning, execution, and reflection through measurable sensor feedback.
ACT-style policies with privileged simulation information and real-world demonstrations.
View videoLiDAR-based cone detection, occupancy-grid mapping, and TEB planning for autonomous racing.
View videoA multimodal perception pipeline that turns scene understanding into executable object poses.
🏆Awards
National College Student Intelligent Vehicle Competition, ROS Group · National Second Prize
National Scholarship
National College Student Mathematical Modeling Competition · National Second Prize
Interdisciplinary Contest in Modeling · Finalist
🌿Beyond Research
I enjoy photography, outdoor exploration, music, hardware tinkering, and electronics. These interests keep me close to the physical world — the same world where robots eventually have to act.