Linpeng Peng
Papers
5
Total Citations
29
H-Index
3
About
Linpeng Peng is a robotics researcher whose work sits at the intersection of computer vision, deep learning, and autonomous manipulation. His research focuses primarily on robotic grasping, place recognition, and pose estimation — core challenges that must be solved for robots to operate effectively in unstructured, real-world environments. Peng's most influential contribution, "Ensemble Bootstrapped Deep Deterministic Policy Gradient for Vision-Based Robotic Grasping" (2021, 19 citations), tackles the fundamental limitation of manipulators that can only grasp pre-specified objects. By leveraging deep reinforcement learning without relying on prior knowledge, he moves robots closer to the adaptable, generalizable grasping ability humans develop through experience. His subsequent work on LiteGrasp further advances this domain by introducing semi-supervised knowledge distillation to achieve efficient grasp detection with reduced dependence on large labeled datasets — a practical step toward real-world deployment. Beyond grasping, Peng has contributed meaningfully to robot localization through his coarse-to-fine place recognition framework combining attention-guided descriptors with overlap estimation, and to industrial automation via real-time 6D pose estimation for bin picking. Together, his publications reflect a consistent commitment to making robotic perception faster, smarter, and more practically deployable across manufacturing and autonomous systems applications.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5