The data sets were collected during motorcycle trips near Vienna in 2021 and 2022. The behavior was split into different classes using videos (not part of the published data due to privacy concerns) and then cut into segments of 10 seconds.
Context and methodology
The data set was collected to show how accurate motorcyclist behavior can be assessed using IMU data
The work follows the ideas published in http://hdl.handle.net/20.500.12708/43982
The authors have a background in geodesy and computer science respectively and work in the field of geoinformation / navigation
Technical details
The data are stored as CSV files
Each file contains data from a unique behavior and has a length of 10 seconds
Each file has a header describing the columns
Units for acceleration are meters per squared second, units for angles are degrees
The files are names AB_Daten_D_C.csv
D: Datum of the trip (as YYYY_MM_DD)
A: Behavior (cruise, fun, overtake, traffic, or wait)
B: Number of the occurrence of this behavior during the trip
C: Number of the segment within the occurrence
The files are grouped by folders named after the corresponding behavior
The IMU used to collect the data was a XSENS MTi