About

Puze Liu is a robotics researcher whose work sits at the intersection of motion planning, safe reinforcement learning, and real-time robot control. His research addresses one of the field's most pressing challenges: enabling robots to operate safely, efficiently, and intelligently in dynamic, real-world environments alongside humans. Liu's most influential contribution, "Regularized Deep Signed Distance Fields for Reactive Motion Generation" (2022, 32 citations), demonstrates his expertise in collision-aware planning, using neural representations to enable real-time obstacle avoidance. This work complements his broader portfolio on constraint manifold methods — a recurring theme across multiple papers — where he develops frameworks for kinodynamic planning and reinforcement learning that respect physical and safety constraints during both training and deployment. His work on safe reinforcement learning (17 citations) and high-speed robot air hockey (14 citations) highlights his talent for bridging theoretical frameworks with demanding real-world applications. More recently, Liu has pushed into embodied AI, developing ROS-LLM, a framework integrating large language models with robot operating systems, signaling a bold expansion into natural language-driven robotics. With over 100 cumulative citations and publications spanning 2021 to 2026, Liu represents an emerging voice shaping the future of intelligent, constraint-aware, and human-collaborative robotic systems.

Research Focus

Key Achievements

6
H-Index
10
Papers
106
Total Citations
11
Avg Citations/Paper
🏆 Most Cited Paper
Regularized Deep Signed Distance Fields for Reactive Motion Generation
32 citations · 2022
📈 Most Prolific Year: 2023 (3 Papers)
🤝 Key Collaborators: 44
🏛 Institutions: Technische Universität Darmstadt, Laboratoire d'Informatique de Paris-Nord, German Research Centre for Artificial Intelligence, ETH Zurich

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 · 14 days ago