AssemblyHAR
Khodadadi, Atieh
1; Reimann de la Cruz, Nicolas
1 Institut für Angewandte Informatik und Formale Beschreibungsverfahren (AIFB), Karlsruher Institut für Technologie (KIT)
1; Reimann de la Cruz, Nicolas 1 Institut für Angewandte Informatik und Formale Beschreibungsverfahren (AIFB), Karlsruher Institut für Technologie (KIT)
Abstract (englisch):
AssemblyHAR provides wrist-worn triaxial accelerometer data recorded from 16 participants during 26 sessions of two manual assembly processes. The release includes raw and preprocessed sensor data, activity annotations, BLE connection logs, metadata, XES event logs, and windowed datasets for human activity recognition and process mining.
| Zugehörige Institution(en) am KIT | Institut für Angewandte Informatik und Formale Beschreibungsverfahren (AIFB) |
| Publikationstyp | Forschungsdaten |
| Publikationsdatum | 06.10.2026 |
| Erstellungsdatum | 20.12.2025 |
| Identifikator | DOI: 10.35097/46ntdksar70pfmmn KITopen-ID: 1000197569 |
| Lizenz | Creative Commons Namensnennung 4.0 International |
| Projektinformation | ONE4ALL (EU, EU 9. RP, 101091877) |
| Schlagwörter | Human activity recognition; Wearable sensors; Accelerometer; Manual assembly; Manufacturing; Process mining; Event logs; Process data |
| Liesmich | AssemblyHAR is organized into two main directories: Raw_Data and Processed_Data. Raw_Data contains the original accelerometer recordings, button-based annotation logs, and BLE connection logs for 26 experimental sessions. Processed_Data contains labelled and preprocessed sensor recordings, validation reports, summary statistics, visualizations, XES event logs, and windowed datasets with grouped and ungrouped activity labels. The windowed datasets include predefined training, validation, and test partitions. A detailed description of the directory structure, file formats, activity labels, and data processing is provided in the accompanying README.md file. |
| Art der Forschungsdaten | Dataset |
| Referent/Betreuer | Lazarova-Molnar, Sanja |