Radu Corcodel
Papers
9
Total Citations
60
H-Index
5
About
Radu Corcodel is a roboticist whose research sits at the intersection of manipulation, tactile perception, and learning from human demonstration. His work tackles some of the hardest open problems in robotics: enabling robots to perform long-horizon assembly tasks, interact with partially observable environments, and generalize skills across novel objects. His most influential paper, “Interactive Planning Using Large Language Models for Partially Observable Robotic Tasks” (2024), has already garnered 23 citations, showcasing the community’s interest in his approach to combining LLMs with interactive planning for open-vocabulary manipulation. Corcodel has also pioneered the Tactile Ensemble Skill Transfer (TEST) framework, an offline reinforcement learning method that leverages tactile feedback for furniture assembly—a notoriously difficult domain. His contributions extend to contact-implicit trajectory optimization, which eliminates the need for parameter tuning, and to constrained dynamic movement primitives that guarantee safety during skill learning. With additional work on audio-visual scene understanding and interactive tactile perception for novel object classification, Corcodel is building a comprehensive toolkit for autonomous, dexterous robots that can learn from and adapt to the real world.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4Tuning-Free Contact-Implicit Trajectory Optimization5 citations · 2020
- 5
- 6
- 7Autonomous Robotic Assembly: From Part Singulation to Precise Assembly4 citations · 2024
- 8
- 9