Jiyao Zhang
Papers
5
Total Citations
49
H-Index
3
About
Jiyao Zhang is at the forefront of modern robotics, specializing in vision-based manipulation, 3D scene understanding, and robot state estimation. His research tackles the fundamental challenge of enabling robots to perceive and interact with complex, unstructured environments using only RGB cameras. Zhang’s most influential work, *RGBGrasp* (24 citations), introduces a novel method for object grasping by leveraging neural radiance fields to capture multiple views during a robot arm’s movement, eliminating the need for expensive depth sensors. He further advances robot perception with *Robot Structure Prior Guided Temporal Attention* (13 citations), which addresses the difficult problem of online camera-to-robot pose estimation from single-view image sequences, overcoming self-occlusion and visual ambiguity. In *LVDiffusor* (7 citations), Zhang distills functional rearrangement priors from large models into a diffusion-based framework, enabling robots to plan precise, task-oriented object arrangements. His work on *RoboKeyGen* (3 citations) extends this to joint angle and pose estimation via diffusion-based 3D keypoint generation. With a growing citation impact, Zhang is shaping a future where robots perceive and act with unprecedented autonomy and dexterity.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5