Checheng Yu 喻车澄

I'm an 4th year undergraduate student at Nanjing University, majored in Automation in School of Management and Engineering.

I'm currently honored to be a research assistant advised by Prof. Lin Shao at National University of Singapore. Recently, my research has focused on foundation model-based embodied AI and robot skill learning 🤖. Previously, I worked with Dr. Yuxiang Sun and Junjie Zhao, studying game agents 🎮 based on reinforcement learning and large models.

Email  /  Scholar  /  Github  /  WeChat  /  CV

×

Research

I'm interested in Robotics 🤖, Reinforcement learning 🕹️ and Multi-agent system 👥. Most of my research is about decision-making models in different scenarios. I'm also working to develop embodied AI with efficient policy learning system and lifelong learning capability. Feel free to contact me if you want to work with me.

In the the real-world robot environment, the capability of the agent in MAS is tied to the physical composition of the robot. To improve the collaboration among heterogeneous robots with varying embodiment capabilities, we introduced a multi-agent framework EMOS combining robot capabilities with task planning. To evaluate how well our MAS performs, we designed Habitat-MAS benchmark, including four tasks: 1) navigation, 2) perception, 3) manipulation, and 4) comprehensive multi-floor object rearrangement.

We proposes a two-layer agent task planning, issues and executes decision commands through the interaction of natural language, and carries out simulation verification through the wargame simulation environment. Through the game confrontation simulation experiment, the intelligent decision-making ability of the large language model is significantly stronger than the commonly used reinforcement learning AI and rule-based AI.

Education & Experience

National University of Singapore Singapore, 2024

Nanjing University Nanjing, China, 2021-2025

Thanks for your visiting 😊. Do not click this button!!! Button

This website' source code is borrowed from Jonbarron.