Papers
2
Total Citations
27
H-Index
2
About
Yujun Zhu is a robotics researcher whose work sits at the intersection of manipulation planning, reinforcement learning, and safe control. His primary contributions focus on enabling robots to perform complex, unstructured tasks with greater autonomy and reliability. In his highly cited 2022 work, "Manipulation Planning From Demonstration Via Goal-Conditioned Prior Action Primitive Decomposition and Alignment" (18 citations), Zhu tackles the critical challenge of trajectory distribution shift, proposing a novel framework that allows prior action primitives to be effectively adapted to new tasks by leveraging goal-conditioned decomposition and alignment. This work advances the field of learning from demonstration by making hierarchical task structures more transferable. Earlier, in his 2019 paper "Reinforcement Learning for Robotic Safe Control with Force Sensing" (9 citations), Zhu addressed the stability and reliability gaps in reinforcement learning for manipulation, integrating force sensing to improve safety in contact-rich environments. By bridging the gap between hierarchical planning and robust learning-based control, Zhu’s research provides practical pathways for deploying robots in real-world settings where adaptability and safety are paramount.
Research Focus
Key Achievements
Top Papers
- 1
- 2Reinforcement Learning for Robotic Safe Control with Force Sensing9 citations · 2019