Humanoid interaction
Whole-body control and motion imitation for physically grounded interaction between humanoid robots.
I am a master's student at Shanghai Jiao Tong University. My current interests are physically grounded robot behavior, sim-to-real learning, and multi-agent autonomous systems.
Research
I am interested in methods and systems that help embodied agents learn reliable behavior from simulation, data, and physical interaction.
Simulation systems
Synthetic data and realistic simulation pipelines for autonomous agents, with attention to evaluation and generalization.
Collaborative autonomy
Multi-agent perception and communication-efficient systems for autonomous driving and embodied coordination.
Selected Work
A compact selection of research projects and papers. This page is a public overview rather than a full CV.
Preprint
It Takes Two: Learning Interactive Whole-Body Control Between Humanoid Robots
A dual-humanoid motion imitation framework for physically grounded whole-body interaction.
ICCV 2025
Unraveling the Effects of Synthetic Data on End-to-End Autonomous Driving
SceneCrafter studies realistic synthetic driving data through a 3D Gaussian Splatting based simulation pipeline.
TPAMI 2025
Towards Collaborative Autonomous Driving: Simulation Platform and End-to-End System
V2Xverse provides a simulation platform for vehicle-to-everything aided autonomous driving.
CVPR / ICCV
Collaborative Perception and Roadside Simulation
Selected work on communication-efficient collaborative perception, temporal BEV prediction, and roadside foreground simulation.
Contact
I am happy to hear from people working on embodied AI, robot learning, humanoid control, and simulation-to-real systems.