Joschka Boedecker

About

Joschka Boedecker is a prominent researcher at the intersection of deep reinforcement learning, robot navigation, and autonomous systems. His work has fundamentally advanced how robots learn to perceive, adapt, and act in complex real-world environments without relying on explicit mapping or localization. His most celebrated contribution, "Deep Reinforcement Learning with Successor Features for Navigation across Similar Environments" (2017, 283 citations), demonstrated how robots could generalize navigation skills across environments using sensor-based learning alone — a significant step toward practical autonomous robotics. Boedecker has made substantial contributions to bridging the simulation-to-reality gap, notably through his VR-Goggles framework (2019, 107 citations), which enables policies trained in simulation to transfer effectively to real-world visual control tasks. His influential survey on deep network solutions for robotic control (2016, 70 citations) helped establish foundational vocabulary and direction for the field. Beyond navigation, his research spans human-robot interaction, brain-computer interfaces, interactive imitation learning, and affordance-based manipulation. Work such as "Correct Me If I am Wrong" reflects his commitment to sample-efficient, human-in-the-loop learning. Collectively, his contributions have shaped modern robot learning methodology, making him a key figure in intelligent autonomous systems research.

Research Focus

Key Achievements

Top Papers

1. 1
   Deep reinforcement learning with successor features for navigation across similar environments

   283 citations · 2017
2. 2
   VR-Goggles for Robots: Real-to-Sim Domain Adaptation for Visual Control

   107 citations · 2019
3. 3
   A Survey of Deep Network Solutions for Learning Control in Robotics: From Reinforcement to Imitation

   70 citations · 2016
4. 4
   Autonomous Learning of State Representations for Control: An Emerging Field Aims to Autonomously Learn State Representations for Reinforcement Learning Agents from Their Real-World Sensor Observations

   51 citations · 2015
5. 5
   A service assistant combining autonomous robotics, flexible goal formulation, and deep-learning-based brain–computer interfacing

   49 citations · 2019
6. 6
   Correct Me If I am Wrong: Interactive Learning for Robotic Manipulation

   37 citations · 2022
7. 7
   Learning Human-Aware Robot Navigation from Physical Interaction via Inverse Reinforcement Learning

   33 citations · 2020
8. 8
   Deep Reinforcement Learning with Successor Features for Navigation across Similar Environments

   33 citations · 2016
9. 9
   Affordance Learning from Play for Sample-Efficient Policy Learning

   29 citations · 2022
10. 10
    Flexible Coordination of Multiagent Team Behavior Using HTN Planning

    25 citations · 2006

Key Collaborators

Contact & Links