About

Jan Peters is a pioneering researcher at the intersection of machine learning and robotics, whose work has fundamentally shaped how autonomous systems learn and execute complex motor behaviors. Best known for his landmark 2013 survey on reinforcement learning in robotics — now cited over 3,000 times — Peters has spent his career bridging the gap between theoretical machine learning and real-world robotic applications. His foundational contributions to policy gradient methods, including the influential Natural Actor-Critic framework and seminal work on policy gradients for robotics (collectively exceeding 1,300 citations), established core algorithmic tools that researchers worldwide continue to build upon. Peters has also advanced the field through probabilistic movement primitives and motor skill learning, enabling robots to acquire, generalize, and adapt behaviors in unstructured environments — demonstrated memorably through robot table tennis. His surveys on model learning for robot control and imitation learning reflect a commitment to synthesizing and guiding the broader research community. With multiple papers exceeding hundreds of citations and a body of work spanning reinforcement learning, imitation learning, and operational space control, Peters stands as one of the most influential figures in modern robot learning research.

Research Focus

Key Achievements

73
H-Index
360
Papers
22,544
Total Citations
63
Avg Citations/Paper
🏆 Most Cited Paper
Reinforcement learning in robotics: A survey
3,055 citations · 2013
📈 Most Prolific Year: 2014 (25 Papers)
🤝 Key Collaborators: 480
🏛 Institutions: Technische Universität Darmstadt, University of Southern California, Max Planck Institute for Biological Cybernetics, Max Planck Society, Max Planck Institute for Intelligent Systems, Institut des Sciences Cognitives Marc Jeannerod

Top Papers

  1. 1
  2. 2
  3. 3
    Natural Actor-Critic
    751 citations · 2008
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 0 days ago