About

Ziyuan Jiao’s research lies at the intersection of robotics, scene understanding, and manipulation planning, with a focus on enabling autonomous agents to perceive, reason about, and act within complex, interactive environments. A central theme in Jiao’s work is the reconstruction of 3D scenes not merely as geometric models, but as functional, actionable spaces. For instance, their 2021 paper on panoptic mapping and CAD model alignment (30 citations) rethinks scene reconstruction from an embodied agent’s perspective, emphasizing the underlying functions and constraints that provide actionable information. This is complemented by their work on sequential manipulation planning using a novel 3D scene graph representation called contact graph⁺ (cg⁺), which enables efficient reasoning for multi-step tasks. Jiao’s contributions extend to multi-agent systems, as seen in their decentralized prioritized motion planning method for UAVs (30 citations), and to legged locomotion, where they identified dynamic similarities between bipedal and quadrupedal bounding gaits (29 citations). More recently, Jiao has explored closed-loop, open-vocabulary mobile manipulation using large vision-language models like GPT-4V, and part-level scene reconstruction that affords robot interaction. With over 200 total citations and a growing portfolio of work that bridges perception, planning, and control, Jiao is establishing a distinctive research program that treats scene understanding as the foundation for intelligent robot autonomy.

Research Focus

Key Achievements

8
H-Index
11
Papers
204
Total Citations
19
Avg Citations/Paper
🏆 Most Cited Paper
Sequential Manipulation Planning on Scene Graph
34 citations · 2022
📈 Most Prolific Year: 2022 (3 Papers)
🤝 Key Collaborators: 33
🏛 Institutions: Peking University, University of Michigan–Ann Arbor, Beijing Academy of Artificial Intelligence, University of California, Los Angeles, Beijing Institute for General Artificial Intelligence

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