{IMU}-Based Object Activity Recognition: An Industrial Dataset and Deep Learning Approach for Enhanced Process Transparency in Logistics

This work addresses activity recognition of industrial objects, specifically targeting pallet activities in Cyber-Physical Systems ({CPS}) enabled warehouses. Three key contributions are presented: (1) the design and implementation of a specialized pipeline for collecting and processing sensor data of pallet activities in real warehouse scenarios, (2) the first industrially collected dataset capturing pallet activities in an operational warehouse environment, and (3) a {CNN}-{LSTM}-based deep learning classifier, evaluated through an ablation study across different sensor types and a hyperparameter study.

  • Veröffentlicht in:
    2026 IEEE 31st International Conference on Emerging Technologies and Factory Automation (ETFA)
  • Typ:
    Inproceedings
  • Autoren:
    Brandt, Marc Julian; Kuhlmann, Jean Lenard; Foot, Hermann; Krooß, Alexander; Kandemir, Ferdi; Jäkel, Jan Phillipp
  • Jahr:
    2026
  • Source:
    https://ieeexplore.ieee.org/abstract/document/11725132/authors

Informationen zur Zitierung

Brandt, Marc Julian; Kuhlmann, Jean Lenard; Foot, Hermann; Krooß, Alexander; Kandemir, Ferdi; Jäkel, Jan Phillipp: {IMU}-Based Object Activity Recognition: An Industrial Dataset and Deep Learning Approach for Enhanced Process Transparency in Logistics, 2026 IEEE 31st International Conference on Emerging Technologies and Factory Automation (ETFA), 2026, 1--8, September, https://ieeexplore.ieee.org/abstract/document/11725132/authors, Brandt.etal.2026a,

BibTeX-Zitat öffnen