Bewegungsklassifikation mithilfe mobiler Sensoren

Authors

  • Sascha Feldhorst Lehrstuhl für Förder- und Lagerwesen, TU Dortmund
  • Michael ten Hompel Lehrstuhl für Förder- und Lagerwesen, TU Dortmund

DOI:

https://doi.org/10.2195/lj_Proc_feldhorst_de_201605_01

Keywords:

Bewegungsklassifikation, Kommissionieren, Kommissionierung, Kommissionierzeitanteile, Maschinelles Lernen, Mobile Sensoren

Abstract

This constribution introduces a new methodology to analyze the manual order picking process in warehouses. The approach relies on a sensor-based motion classification which is already used in other domains like sports or medical science. Thereby, different sensor data, e. g. accelerations or rotation rates are continuously recorded during the order picking.With help of this data the motion can be analysed to identify different motion states. Therefore, this paper defines all relevant motion states within the order picking process. These states are part of a classification system which works with methodologies from the field of machine learning. The classification is done with a supervised learning approach which achieves an accuracy of up to 78,94 percent.

Downloads

Published

2016-05-27

How to Cite

Feldhorst, S., & ten Hompel, M. (2016). Bewegungsklassifikation mithilfe mobiler Sensoren. Logistics Journal: Proceedings, (11.2). https://doi.org/10.2195/lj_Proc_feldhorst_de_201605_01