Chenjie Yang
Papers
7
Total Citations
145
H-Index
6
About
Chenjie Yang is a leading researcher in autonomous robotic manipulation, with a focus on enabling robots to perceive, reason, and act in complex, multi-object environments. Their work tackles the fundamental challenge of task-oriented grasping in cluttered scenes, where objects are stacked or interact in ways that confuse conventional systems. Yang pioneered the use of multi-task convolutional neural networks to simultaneously identify objects and plan grasps, achieving real-time performance in object stacking scenes—a breakthrough that has garnered over 76 citations. They further advanced the field by introducing CRF-based semantic models and gated graph neural networks to detect visual manipulation relationships, allowing robots to understand not just what to grasp, but in what sequence. This work on relationship detection, with over 20 citations, addresses the critical problem of redundant or missed grasps in multi-object settings. Yang’s research also extends to autonomous tool construction, where robots select and assemble parts to reconstruct tools—a step toward more adaptive and intelligent systems. With a growing citation record and a focus on practical, real-time solutions, Yang is shaping the future of robotic perception and manipulation.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5Autonomous Tool Construction with Gated Graph Neural Network6 citations · 2020
- 6
- 7