Publications * indicates equal contribution | highlighted publications are selected
* indicates equal contribution | highlighted publications are selected
LLM-Raft: Enhancing Urban Traffic Efficiency and Safety through Decentralized Coordination of Autonomous Vehicles
NeurIPS 2025 Workshop on UrbanAI paper
Education
Shanghai Jiao Tong University 2021 – 2025
Electrical Engineering
Experience
Shanghai Qi Zhi Institute 03 / 2025 – Present
Research Assistant
- Focusing on video robot learning and embodied AI
- Developing intelligent systems that integrate robot action policies with generative AI capabilities
- Investigating generalizable representations and high-quality training data generation
EmPRISE Lab, Cornell 07 / 2024 – 12 / 2024
Research Intern
- Developing intelligent assistive robots framework powered by MPC and diffusion policy
- Enabling natural human-robot interaction for transferring task
- Advancing collaborative capabilities between assistive robots
Dequan Wang's Group, SJTU 09 / 2023 – 06 / 2024
Research Intern
- Developing LLM-powered vehicle coordination frameworks for autonomous driving and intelligent transportation scenarios, focusing on multi-agent consensus and traffic optimization
- Developing LLM-powered collaborative frameworks for multi-robot systems, enabling intelligent task coordination and distributed decision-making
Research
Scene and Task Generation Interface for Rob Simulation
11 / 2024 – 12 / 2024
- Developed an interface system for robotic caregiving scene generation, supporting diverse care task configurations and scene layouts
- Integrated an LLM-based task description generator for automatic generation of detailed care task descriptions and scene requirements
- Implemented a scene editor functionality supporting interactive object placement, obstacle addition, and environment parameter customization
- Enabled export of standardized scene description files for direct use in robot simulation environments and algorithm evaluation
Coordinating Assistive Transferring with Multi-Agent MPC
08 / 2024 – 10 / 2024
- Proposed a real-time scene-aware multi-agent MPC architecture for autonomous systems
- Developed a transferable MPC framework for multi-robot collaboration
- Developed an intelligent collaborative framework for assistive transfer tasks by integrating Kinova 6-DOF robotic arm, Hoyer sling, and powered wheelchair systems
LLM-Powered Vehicle Coordination for Autonomous Driving
09 / 2023 – 06 / 2024
- Developed LLM-powered decentralized coordination framework for autonomous vehicles in urban traffic scenarios
- Implemented multi-agent consensus mechanisms for traffic efficiency and safety optimization
- Designed and evaluated coordination protocols for intersection management and lane merging
- Conducted extensive simulations demonstrating improved traffic flow and reduced collision rates
Outreach
Workshop
06 / 2024 – 09 / 2024
- Developed comprehensive simulation frameworks for robot assistive dressing and bathing tasks using RCareWorld platform
- Established complete robot control algorithm evaluation system with automated assessment pipeline via EvalAI integration
- Hosted regular office hours and created comprehensive documentation for 56 registered teams
Reviewer
ACM/IEEE International Conference on Human-Robot Interaction (HRI) 2025, 2026