Zhangli Zhou
Papers
12
Total Citations
245
H-Index
7
About
Zhangli Zhou is an emerging robotics researcher whose work spans two interconnected frontiers: vision-based robotic grasping and formal methods for robot motion planning. Zhou's most influential contribution is TF-Grasp, a transformer-based architecture for robotic grasp detection that leverages local window attention to efficiently capture contextual information — a paper that has garnered 148 citations since 2022, signaling substantial impact in the computer vision and robotics communities. Complementing this, Zhou has pioneered natural language-guided grasping and intuitive human-robot interfaces, including a novel eye-tracking-based system enabling users to direct robotic manipulation through gaze alone. Equally notable is Zhou's rigorous work in temporal logic motion planning, developing frameworks such as extended predicate-based temporal logic (E-pTL) and planning decision trees to enable fast, reactive task execution for multi-robot systems and quadruped robots navigating dynamic, unstructured environments. This body of work directly addresses the practical challenge of real-time adaptability in human-robot collaboration. Zhou has also contributed to unsupervised representation learning for robotic vision, reducing dependency on costly labeled datasets. Across all these directions, Zhou demonstrates a rare ability to bridge theoretical rigor with deployable, human-centered robotic systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5
- 6
- 7Unsupervised Representation Learning for Visual Robotics Grasping7 citations · 2022
- 8
- 9
- 10