Papers
6
Total Citations
190
H-Index
5
About
Peiyuan Ni is a leading researcher in robotic manipulation, with a focus on grasp generation, pose estimation, and assembly in unstructured environments. Their most impactful work, "PointNet++ Grasping: Learning An End-to-end Spatial Grasp Generation Algorithm from Sparse Point Clouds" (2020), has garnered 144 citations and introduces a novel approach that bypasses traditional grasp sampling pipelines by directly generating grasps from sparse point clouds using deep learning. This innovation significantly reduces computational cost and improves efficiency for novel object grasping. Ni further advanced the field with "A New Approach Based on Two-stream CNNs for Novel Objects Grasping in Clutter" (2018), which tackles grasping in cluttered settings, and "Compliant Robotic Assembly based on Deep Reinforcement Learning" (2021), applying reinforcement learning to high-precision industrial assembly tasks like peg-in-hole. More recently, Ni developed "PanelPose: A 6D Pose Estimation of Highly-Variable Panel Object for Robotic Robust Cockpit Panel Inspection" (2023), addressing the challenge of variable textures and point clouds in aerospace inspection. With over 190 total citations, Ni’s work bridges simulation and real-world application, driving progress in autonomous robotics for manufacturing and inspection.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4Compliant Robotic Assembly based on Deep Reinforcement Learning6 citations · 2021
- 5
- 6