Papers
10
Total Citations
106
H-Index
6
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
Top Papers
- 1Regularized Deep Signed Distance Fields for Reactive Motion Generation32 citations · 2022
- 2
- 3
- 4Efficient and Reactive Planning for High Speed Robot Air Hockey14 citations · 2021
- 5ROS-LLM: A Framework for Embodied AI7 citations · 2025
- 6Robot Reinforcement Learning on the Constraint Manifold6 citations · 2021
- 7
- 8
- 9
- 10