Mobile Robotik

Visual Navigation

With the availability of affordable 3D cameras, most prominently the Microsoft Kinect, visual navigation has become increasingly important. At the ITI, we are primarily interested in the application of visual navigation and most importantly visual simultaneous localization and mapping methods to industrial mobile robots. Our focus is therefore on

  • scalability
  • realtime capability
  • robustness

of the navigation algorithms.

Datasets and Source Code

We have recently established our own set of test sequences for the evaluation of large-scale Visual Simultaneous Localization and Mapping. We have further made our VSLAM solutions available on github. If you are interested in the dataset or source code, you will find a more detailed description and further instructions under the following link:

http://www.iti.uni-luebeck.de/forschung/mobile-robotik/visual-navigation/dataset-and-source-code.html

Publikationen

[HSM13] Hartmann, J.;Stechele, W.; Maehle, E.: Self-Adaptation for Mobile Robot Algorithms Using Organic Computing Principles. Architecture of Computing Systems (ARCS), 232-243, Springer Berlin Heidelberg, Prag 2013 [Abstract] [Paper]
[KHF13] Klüssendorff, Jan Helge; Hartmann, Jan; Forouher, Dariush; Maehle, Erik: Graph-Based Visual SLAM and Visual Odometry using an RGB-D Camera. 9th International Workshop on Robot Motion and Control, RoMoCo 13, Wasowo 2013
[HFL12] Hartmann, J.; Forouher, D.; Litza, M.; Klüssendorff, J.H.; Maehle, E.: Real-Time Visual SLAM Using FastSLAM and the Microsoft Kinect Camera. 7th German Conference on Robotics (ROBOTIK 2012), 458-463, München 2012 [Abstract]