Papers
6
Total Citations
189
H-Index
6
About
Marijn Stollenga is a leading researcher in autonomous robotics, with a focus on enabling humanoid robots to learn complex behaviors through intrinsic motivation and curiosity-driven exploration. His work bridges artificial intelligence, robotics, and cognitive science, pioneering methods that allow robots to acquire sensorimotor skills from high-dimensional, unprocessed visual inputs without explicit supervision. Stollenga’s most influential contribution is the development of curiosity-driven reinforcement learning frameworks, such as those presented in his highly cited 2014 paper (75 citations), which demonstrated how artificial curiosity can guide motion planning on humanoid platforms. He further advanced this paradigm with SKILLABILITY (2014), a system that autonomously learns multiple skills as a byproduct of exploration. His work on Task-Relevant Roadmaps (TRMs) and Natural Gradient Inverse Kinematics (NGIK) (2013, 20 citations) introduced a flexible, sampling-based optimizer for complex motion planning. Stollenga also contributed to the Modular Behavioral Environment (MoBeE, 2012), a tightly integrated control system for humanoid vision, action, and reaction. With over 150 total citations, his research continues to inspire new approaches to lifelong learning and autonomous skill acquisition in robotics.
Research Focus
Key Achievements
Top Papers
- 1Curiosity Driven Reinforcement Learning for Motion Planning on Humanoids75 citations · 2014
- 2
- 3Task-relevant roadmaps: A framework for humanoid motion planning20 citations · 2013
- 4The Modular Behavioral Environment for Humanoids and other Robots (MoBeE)18 citations · 2012
- 5
- 6Explore to see, learn to perceive, get the actions for free: SKILLABILITY11 citations · 2014