Garage Positioning Dataset

Cite this
×

Citation

Van de Wynckel, M. (2025). Garage Positioning Dataset. Kaggle. https://doi.org/10.34740/KAGGLE/DS/6654647

@misc{vandewynckel2025kaggle,
	author = {Van de Wynckel, Maxim},
	doi = {10.34740/KAGGLE/DS/6654647},
	year = {2025},
	publisher = {Kaggle},
	title = {Garage {Positioning} {Dataset}},
	url = {https://www.kaggle.com/ds/6654647},
}

Authors

Abstract

  • For this dataset we split up a room in 45 points (with centimetre coordinates in our dataset).
  • The room has the following dimensions: 500cm * 350cm
  • For each point, we record the dataset at 60cm height.
  • For each point, we perform data collection in four directions (indicated with the column ORIENTATION). This means the total amount of data points collected is 45 * 4.
  • Each set of data (every direction for every point) takes approximately 30 seconds
  • We use a smartphone to collect the data (Samsung S20 FE)
  • Our mobile phone collects information about the accelerometer (ACC_), gyroscope (RRATE_), magnetometer (MAG_) and computed pitch, roll, yaw.
  • Our mobile phone also collects the received signal strength (RSSI) of 4 Bluetooth beacons and all Wi-Fi access points in range (denoted using WLAN_ as columns).
  • If an access point was not in range, its value was set to 100 (NOTE: 100 RSSI is invalid, it should always be negative)
  • The Bluetooth beacons have the following coordinates (x, y, z):
    • Beacon 1: (0, 0, 150)
    • Beacon 2: (450, 0, 150)
    • Beacon 3: (250, 316, 150)
    • Beacon 4: (0, 316, 150)
  • Raw results are unprocessed apart from renaming the WLAN access points MAC addresses to pseudonymise the results.
  • 66 WLAN access points were detected