Spatial Perception
Robust mapping, localization, and spatial understanding for real-world robotic systems.
Embodied Spatial AI Lab
Hong Kong University of Science and Technology (Guangzhou)
Building next-generation embodied spatial intelligence systems for robots that perceive, reason, and act in complex real-world environments.
About
Spatial intelligence for embodied robots
EmSAIL Group focuses on next-generation embodied spatial intelligence systems for robotics operating in complex real-world environments. Our goal is to integrate spatial perception, representation, reasoning, and action into a unified embodied intelligence pipeline.
6Research directions
AI + RoboticsCore focus
HKUST(GZ)Location
Research
Research Directions
Embodied Models
Predictive models that support planning, reasoning, and autonomous decision-making.
Robot Policy Learning
Learning generalizable robot skills from demonstrations, interaction, and simulation.
People
Team
Principal Investigator
Prof. Yan Li
Principal Investigator
Research interests: Embodied Spatial AI, Robot Learning, 3D Vision
Researchers & Students
C. Lyu
Chengyu Lyu
Research Assistant
Research interests: Robot Policy LearningS. Feng
Shuo Feng
Research Assistant
Research interests: Vision-Language-Navigation (VLN)Y. Zhao
Yong ZHAO
Visiting Master Student
Research interests: Gaussian Splatting, SLAMC. LIU
Chuan LIU
Visiting Master Student
Research interests: Gaussian Splatting, SLAMPublications
Selected Publications
2026
Publication list coming soon
Add paper title, venue, authors, PDF, code, and project page links here.
Projects
Research Projects
Embodied Spatial Mapping
Unified mapping and representation for embodied agents.
Robot Foundation Models
Vision-language-action models for general-purpose robot control.
Simulation-to-Real Learning
Reliable transfer from virtual training to physical robots.
News
Latest Updates
EmSAIL Lab website launched.
We are recruiting PhD students, MPhil students, research assistants, and interns.
Join & Collaboration
We are recruiting motivated students and collaborators.
Students with strong backgrounds in robotics, computer vision, machine learning, spatial intelligence, or embedded systems are encouraged to apply.