Chengshu Li

Stanford University, Stanford Medicine

Papers

10

Total Citations

456

H-Index

8

About

Chengshu Li is a leading researcher in embodied AI and robot learning, whose work centers on creating the simulation infrastructure and learning algorithms that enable robots to perform complex, everyday household tasks. He is the principal architect of the iGibson simulation environments (iGibson 1.0 and 2.0, with over 190 combined citations), which have become foundational tools for the field by providing large-scale, realistic, and fully interactive home scenes populated with rigid and articulated objects. These environments directly address the critical gap between simulated and real-world robotics. Li’s major contributions extend to algorithm design, most notably through ReLMoGen (over 120 citations), which innovatively integrates motion generation with reinforcement learning to solve challenging mobile manipulation tasks by lifting the action space to high-level subgoals. He has also shaped the evaluation of socially-aware robots, co-authoring key principles and guidelines for social robot navigation (over 70 citations). His recent work includes the massive BEHAVIOR-1K benchmark, a human-centered, 1,000-activity benchmark that grounds robotic tasks in real human desires. Through these efforts, Li has provided the community with the tools, tasks, and evaluation standards necessary to advance toward truly capable household robots.

Research Focus

Key Achievements

8
H-Index
10
Papers
456
Total Citations
46
Avg Citations/Paper
🏆 Most Cited Paper
iGibson 1.0: A Simulation Environment for Interactive Tasks in Large Realistic Scenes
132 citations · 2021
📈 Most Prolific Year: 2021 (3 Papers)
🤝 Key Collaborators: 79
🏛 Institutions: Stanford University, Stanford Medicine

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 13 days ago