Überblick

Open Access

• false

Peer Reviewed

• true

Abstract

• This paper addresses the problem of dealing with different kinds of dynamic obstacles influencing a place recognition task. We improve an existing approach using independent Marcov chain grid maps (iMac). Furthermore, we add a fuzzy classification to exploit the iMac estimation to refine the likelihood field estimation. We can show that the proposed method increases the performance of place recognition, while still being a compact, interpretable framework.

authors

• Bahrmann, Frank
• Böhme, Hans-Joachim
• Hellbach, Sven
• Keil, Sabrina

Veröffentlichungszeitpunkt

• Januar 1, 2014

veröffentlicht in

• Neural Information Processing, pp. 553-562  Zeitschrift