We have organised a workshop during the ECMR 2021 conference with the title “How good is your map? Robotic Map Assessment and Standardisation”. See https://goodmap21.github.io
The workshop was quite well attended, with a wide cast of invited speakers from academia and industry:
- Giorgio Grisetti, Sapienza University of Rome
“SLAM: what still bugs me” - Magnus Lindhé, Electrolux
“A consumer robot perspective on map quality” - Max Pfingsthorn, OFFIS
“Can we make more accurate and actionable semantic maps with human and tactile input?” - Naoki Akai, Nagoya University
“Detection of localization failures with probabilistic modeling” - Francesco Amigoni, Politecnico di Milano
“Standardizing robot map representations” - Mirco Colosi, Bosch
“Standardizing SLAM: exploiting recurrent patterns for modularity and behavioral robustness” - Tomasz Kucner, Örebro University
“The good, the bad, the ugly: Reference-free map quality assessment”
We also had lively group discussions brainstorming on the following questions:
- What characterises a good map, and how to assess that?
- How to present information encoded in a robotic map in a human readable format?
- What is the basic spatial information that should be encoded in a (standardised) map?
- What domains/industries are in need of map standards and map quality assessment?
Check recordings of the invited talks here: https://youtube.com/playlist?list=PLetlvHhgg7yY5EbHMX96WZTFVIGeoUW_P