Papers
15
Total Citations
392
H-Index
8
About
Andras Kupcsik is a leading researcher in robotics, specializing in data-efficient robot learning, contextual policy search, and human-robot interaction. His most influential work, "Data-Efficient Generalization of Robot Skills with Contextual Policy Search" (2013, 91 citations), introduced a hierarchical approach enabling robots to generalize controllers across varying contexts—such as changes in environment or task objectives—without requiring extensive retraining. This foundational contribution, further developed in his 2014 model-based extension (79 citations), has significantly advanced the practical deployment of adaptive robotic systems. Kupcsik has also made notable strides in underwater manipulation, with his 2018 paper (67 citations) streamlining remotely operated vehicle operations by reducing crew requirements and improving efficiency. His work on learning dynamic robot-to-human object handover from human feedback (2017, 58 citations) and forceful manipulation skills from multi-modal demonstrations (2021) underscores his commitment to intuitive, human-centric robot programming. Additionally, his research on modular policies (2014) and the kernel Kalman rule (2019) demonstrates theoretical depth. With over 370 total citations, Kupcsik’s contributions bridge theory and application, from industrial assembly (e-Bike motor assembly, 2023) to service robot benchmarking, making him a key figure in advancing flexible, real-world robotic manipulation.
Research Focus
Key Achievements
Top Papers
- 1Data-Efficient Generalization of Robot Skills with Contextual Policy Search91 citations · 2013
- 2
- 3
- 4Learning Dynamic Robot-to-Human Object Handover from Human Feedback58 citations · 2017
- 5Learning Forceful Manipulation Skills from Multi-modal Human Demonstrations22 citations · 2021
- 6Learning modular policies for robotics20 citations · 2014
- 7The kernel Kalman rule15 citations · 2019
- 8
- 9
- 10