Paperdoi:10.3390/s20226479

Accelerometer-Based Fall Detection Using Machine Learning: Training and Testing on Real-World Falls

Venue listed2020Added Nov 17, 2026

  • LP
  • JK
  • CB
  • LC

Authors: Luca Palmerini, Jochen Klenk, Clemens Becker, Lorenzo Chiari

Sensors 20(22):6479 (MDPI, 2020-11-13), open access

TL;DR

143 real-world falls from 40 subjects (mean age 69.2) in FARSEEING, lower-back inertial sensor at 100 Hz, plus 180 h of ADL recordings (15 subjects × 12 h). Subject-based 5-fold CV. Best method (SVM + multiphase-fall-model features): sensitivity >80%, 0.56 false alarms per hour, F-measure 64.6%.

  • Form factor: lower back

Abstract

Falling is a significant health problem. Fall detection, to alert for medical attention, has been gaining increasing attention. Still, most of the existing studies use falls simulated in a laboratory environment to test the obtained performance. We analyzed the acceleration signals recorded by an inertial sensor on the lower back during 143 real-world falls (the most extensive collection to date) from the FARSEEING repository. Such data were obtained from continuous real-world monitoring of subjects with a moderate-to-high risk of falling. We designed and tested fall detection algorithms using features inspired by a multiphase fall model and a machine learning approach. The obtained results suggest that algorithms can learn effectively from features extracted from a multiphase fall model, consistently overperforming more conventional features. The most promising method (support vector machines and features from the multiphase fall model) obtained a sensitivity higher than 80%, a false alarm rate per hour of 0.56, and an F-measure of 64.6%. The reported results and methodologies represent an advancement of knowledge on real-world fall detection and suggest useful metrics for characterizing fall detection systems for real-world use.

Verbatim from the publisher's DOI record (opens in a new tab) · Publisher licence: https://creativecommons.org/licenses/by/4.0/

Results on ChatPPG leaderboards

All leaderboards

No results from this paper on ChatPPG leaderboards yet. Browse the leaderboards.

Tasks & signals

Signals

Keywords

Fall DetectionReal WorldFARSEEINGFalse Alarm RateSVMLower Back

How it works

How Method works: Setup → Method → TargetSetup: Signals (ACC); Form factor (lower back). Method: Method.SetupSignals: ACCxyzSignalsACCForm factor: lower backForm factorlower backMethodMethod

Cite this paper

Cite this paper with the fields recorded here. Check it against the publisher before submitting.

Palmerini, L., Klenk, J., Becker, C., & Chiari, L. (2020). Accelerometer-Based Fall Detection Using Machine Learning: Training and Testing on Real-World Falls. Sensors , open access, 20(22), 6479. https://doi.org/10.3390/s20226479
Palmerini, Luca, Jochen Klenk, Clemens Becker, and Lorenzo Chiari. 2020. "Accelerometer-Based Fall Detection Using Machine Learning: Training and Testing on Real-World Falls." Sensors , open access 20 (22): 6479. https://doi.org/10.3390/s20226479.
Palmerini, L., Klenk, J., Becker, C. and Chiari, L. (2020) 'Accelerometer-Based Fall Detection Using Machine Learning: Training and Testing on Real-World Falls', Sensors , open access, 20(22), p. 6479. doi: 10.3390/s20226479.
Palmerini, Luca, et al. "Accelerometer-Based Fall Detection Using Machine Learning: Training and Testing on Real-World Falls." Sensors , open access, vol. 20, no. 22, 2020, p. 6479. https://doi.org/10.3390/s20226479.
L. Palmerini, J. Klenk, C. Becker, and L. Chiari, "Accelerometer-Based Fall Detection Using Machine Learning: Training and Testing on Real-World Falls," Sensors , open access, vol. 20, no. 22, p. 6479, 2020, doi: 10.3390/s20226479.
@article{palmerini2020accelerometerbased,
  title         = {{Accelerometer-Based Fall Detection Using Machine Learning: Training and Testing on Real-World Falls}},
  author        = {Palmerini, Luca and Klenk, Jochen and Becker, Clemens and Chiari, Lorenzo},
  year          = {2020},
  journal       = {Sensors , open access},
  volume        = {20},
  number        = {22},
  pages         = {6479},
  doi           = {10.3390/s20226479}
}

Added to the ChatPPG research hub