Stefan Klanke

About

Stefan Klanke is a robotics and machine learning researcher whose work sits at the intersection of adaptive control, probabilistic learning, and anthropomorphic robotic systems. His research has made significant contributions to how robots learn and execute complex motor tasks, drawing inspiration from biological principles of movement and control. Klanke's most influential work, "Multi-task Gaussian Process Learning of Robot Inverse Dynamics" (2008, 91 citations), demonstrated how probabilistic machine learning techniques could be leveraged to teach robots to compute joint torques across varying load conditions — a critical challenge in adaptive robotic control. His subsequent research extended these ideas into variable impedance actuation, exploring how anthropomorphic robots could achieve the agility, robustness, and efficiency characteristic of biological systems through stochastic optimization and optimal feedback control frameworks. A recurring theme in his work is learning control policies from constrained or variable data — particularly situations where environmental constraints are unobservable or context-dependent, mirroring the challenges humans face in everyday motor tasks. His path planning work for redundant 7-DOF robotic arms further demonstrates his breadth across motion planning and real-time robotics. With contributions appearing at leading venues including ICRA and IROS, Klanke's research has meaningfully advanced the field of robot learning and biologically inspired control, accumulating over 300 citations across his key publications.

Research Focus

Key Achievements

Top Papers

1. 1
   Multi-task Gaussian Process Learning of Robot Inverse Dynamics

   91 citations · 2008
2. 2
   Learning impedance control of antagonistic systems based on stochastic optimization principles

   47 citations · 2010
3. 3
   Dynamic Path Planning for a 7-DOF Robot Arm

   33 citations · 2006
4. 4
   A novel method for learning policies from variable constraint data

   32 citations · 2009
5. 5
   IROS 2009. IEEE/RSJ International Conference on Intelligent Robots and Systems, 2009

   32 citations · 2009
6. 6
   Optimal Feedback Control for anthropomorphic manipulators

   26 citations · 2010
7. 7
   ICRA '09. IEEE International Conference on Robotics and Automation, 2009

   19 citations · 2009
8. 8
   Exploiting sensorimotor stochasticity for learning control of variable impedance actuators

   16 citations · 2010
9. 9
   Learning potential-based policies from constrained motion

   13 citations · 2008
10. 10
    A novel method for learning policies from constrained motion

    8 citations · 2009

Key Collaborators

Contact & Links