Uni-Logo
You are here: Home Research
Artikelaktionen

Research

Research Focus

The focus of our research lies on robots acting in human environments. First, we investigate the interaction between robots and non-expert users and aim at designing intuitive programming interfaces. Second, we are especially interested in robot navigation in challenging environments. For example, we develop techniques for humanoid robots navigating in multi-level and multi-room indoor environments that contain articulated and movable objects. We develop probabilistic methods for 3D environment modeling as well as for manipulation and navigation. This work is carried out within the SFB/TR8 Spatial Cognition funded by the German Research Foundation and the projects First-MM, ROVINA, and SQUIRREL funded by the EU.

As part of the research training group on Embedded Microsystems funded by the German Research Foundation, we aim at developing effective methods for systems with limited hardware resources. We contribute robust techniques for the interpretation of sensor data and for navigation control of robots with only noisy sensor and actuator devices.

Furthermore, we carry out research within the cluster of excellence BrainLinks-BrainTools. Here, one focus lies on the autonomous execution of manipulation tasks for paralyzed patients based on brain-machine interfaces. We will also analyze human motions for real-time imitation of the recorded motions on humanoid robots and for neurological patients suffering from motion malfunctions. The goal is to improve the patients behavior using closed-loop deep brain stimulation. Both projects are based on close, interdisciplinary collaborations.

Research Topics

  Navigation Through Clutter
  Whole-Body Motion Planning
  Imitation of Human Whole-Body Motions
  Autonomous Navigation Based on Depth Camera Data
  Climbing Complex Staircases and Traversing Ramps
  Efficient Vision-based Obstacle Avoidance
  Efficient Path Planning for Humanoids
  6D Robot Localization in Complex Indoor Environments
  Learning Reliable and Efficient Navigation with a Humanoid
  Metric Localization with Scale-Invariant Visual Features using Monocular Vision
  Learning Efficient Policies for Vision-based Navigation
  Robust Recognition of Complex, Parameterized Gestures
  Multimodal Interaction between a Humanoid Robot and Humans
  Utilizing Learned Motion Patterns to Predict Positions of People
  Adapting Navigation Strategies Using Motion Patterns of People
  Learning Motion Patterns of People
  Prioritized Multi-robot Path Planning
Benutzerspezifische Werkzeuge