Yuanjie Lu

Yuanjie Lu

PhD Candidate at George Mason University

Short Bio

Yuanjie Lu is a PhD candidate in Computer Science at George Mason University, advised by Prof. Xuesu Xiao in the RobotiXX Lab. He is also affiliated with the Center for Human-AI Innovation in Society. His research focuses on embodied AI and robot learning, with an emphasis on foundation models, world models, and deep reinforcement learning for autonomous navigation and control. He develops methods that integrate LLM/VLM-based reasoning, world models, and deep reinforcement learning for robot planning and control in complex environments. His research includes applications to wheeled and legged navigation, as well as humanoid locomotion. He is seeking industry positions as a Research Scientist or Machine Learning Engineer in embodied AI, robot learning, and foundation models.

đź“„ Download Full CV (PDF)

Research Interests

  • Foundation Models for Embodied AI: LLM/VLM-based reasoning, planning, and adaptation for autonomous robots
  • World Models for Embodied Intelligence: Learning predictive representations of the physical world for action-conditioned prediction, planning, and control
  • Deep Reinforcement Learning for Robot Control: Adaptive and hierarchical control, policy learning, and sim-to-real transfer
  • Machine Learning-Augmented Motion Planning: Integrating learned models, planning memory, and classical planners for efficient, executable motion

Collaborations

His research journey has progressed from machine learning and deep learning for autonomous systems, through motion planning algorithms in robotics, to real-world robot navigation and foundation model-driven autonomy.

CollaboratorTopic
Dr. Dongqi Han and Dr. Dongsheng Li
Microsoft Research Asia (MSRA)		Dynamics-aware navigation for quadruped robots
Prof. Chengzhi Mao
Rutgers University		Reliable multimodal reasoning and trustworthy foundation models
Prof. Xiaomin Lin
University of South Florida		LLM/VLM-driven robot navigation
Zhenyu Yang
Chief Researcher, OPPO Xiaobu Assistant		Natural language processing, large language models, and conversational AI
Prof. Tinoosh Mohsenin
Johns Hopkins University		Quadruped navigation
Prof. Nick Hawes
Oxford University		Robot dynamics and adaptive control
Prof. Erion Plaku
U.S. National Science Foundation (NSF), CISE		Motion planning that integrates learning with classical planners
Virginia Transportation Research CouncilGraph neural networks for traffic flow forecasting under work zones and lane closures
Prof. Amarda Shehu and Prof. David Lattanzi
George Mason University		Data-driven anomaly forecasting for autonomous systems

Experiences

Microsoft Research Asia
Research Intern
May–August 2026
Johns Hopkins University
Research Engineer Intern
May–August 2025
Unitree Robotics Ă— RobotiXX Lab
Digital Innovation Summer Experiential Fellow
May–August 2024
University of Oxford
Remote Researcher
Nov 2024 - May 2025
Virginia Department of Transportation (VDOT)
Summer Research Assistant · VTRC-sponsored project
May–August 2020

Recent Publications & Projects

Learning Scene-Aware Humanoid Locomotion through 3D Clutter from Immersive Human Demonstrations (Submitted to ICRA 2027)

Learning scene-aware humanoid locomotion from immersive VR demonstrations, enabling a Unitree G1 to crawl through low passages and navigate narrow gaps.

Scene-aware humanoid locomotion in simulation Unitree G1 navigation demonstration

Quadruped Navigation in Dense Obstacle Environments (Summer 2026 @ Microsoft Research Asia)

Quadruped navigation demonstrations in an indoor corridor with obstacles.

APPLV: Adaptive Planner Parameter Learning from Vision-Language-Action Model (IROS 2026)

This work studies how vision–language models (VLMs) can be leveraged to adaptively adjust classical motion planner parameters based on semantic and contextual understanding of the environment

Moving Through Clutter: Scaling Data Collection and Benchmarking for 3D Scene-Aware Humanoid Locomotion via Virtual Reality

Generates scenes with controllable clutter levels and captures embodiment-consistent, whole-body human motion through immersive VR navigation

CORAL: Contextual Reasoning And Local Planning in A Hierarchical VLM Framework for Underwater Monitoring (IROS 2026)

a framework that decouples high-level semantic reasoning from low-level reactive control for underwater monitoring

Adaptive Dynamics Planning (ADP) for Robot Navigation (ICRA 2026)

Combining Motion Planning with TD3-Based Reinforcement Learning for Real-Time Dynamics Adaptation in Mapless, Constrained Environments

Autonomous Navigation for Legged Robots (Summer 2025 @ JHU)

Autonomous Navigation for Legged Robots in Complex Environments

News