Learning for Localization and Mapping
Workshop at IROS 2017
September 28th, 14:00 - 17:30
Room 211-214

The goal of this workshop is to present and discuss developments in learning-based approaches for localization and mapping systems. We aim at bridging the gap between robotics and machine learning research by bringing together researchers from both fields. This workshop focuses on developing a clear vision for future directions of research of learning-based mapping techniques and novel machine learning algorithms that promise to further bolster localization and mapping systems.

 

We are soliciting papers on any potential topic within the scope of the workshop. The papers can be submitted in IROS format (4-6 pages, references may go on an extra page). Accepted papers will be presented in a poster session. The pdf of the posters alongside with the abstracts will be published on the workshop website. Please refer to the submission page for further instructions.


The topics of interest involve but are not limited to:

  • Improving distinctiveness of place representations for place recognition and loop closure.

  • Learning-based approaches for map maintenance and map compression.

  • Selecting useful landmarks for map creation.

  • Multi-modal landmark descriptors and matching between different sensor types.

  • Representation of objects as landmarks in SLAM.

  • Deployment of learning-based approaches on resource constrained robotic platforms.

  • Uncertainty quantification for deep learning and its applicability in a Bayesian fusion setting.

  • Success stories and lessons learned (including negative results).


Invited Speakers
We are proud to have a group of famous invited speakers presenting at our workshop involving
  • Wolfram Burgard, University of Freiburg
  • Jana Kosecka, George Mason University
  • Stefan Leutenegger, Imperial College London
  • Simon Lynen, Google 
  • Marc Pollefeys, Microsoft 
  • Fabio Ramos, University of Sydney