Related Experiment Video
Updated: Nov 10, 2025
08:05
Design and Analysis for Fall Detection System Simplification
Published on: April 6, 2020
11.0K
AnkFall-Falls, Falling Risks and Daily-Life Activities Dataset with an Ankle-Placed Accelerometer and Training Using
Francisco Luna-Perejón1,2, Luis Muñoz-Saavedra1,2, Javier Civit-Masot1,2
1Architecture and Computer Technology Department, ETSII-EPS, University of Seville, 41004 Sevilla, Spain.
Sensors (Basel, Switzerland)
|April 3, 2021
Summary
This study introduces a novel ankle-worn device for elderly fall detection, addressing discomfort and forgetfulness associated with traditional wearables. The developed dataset and initial deep learning model demonstrate the system's feasibility for remote monitoring.
Area of Science:
- Gerontology
- Biomedical Engineering
- Data Science
Background:
- Falls represent a major cause of injury and disability in the elderly population.
- Existing remote monitoring systems often cause user discomfort and non-compliance due to device placement (waist/wrist).
- Pandemic scenarios highlight the need for remote health monitoring solutions that minimize unnecessary mobility.
Purpose of the Study:
- To design and collect a new dataset for fall detection and activities of daily living (ADLs) using an ankle-placed device.
- To overcome the limitations of current wearable fall detection systems, focusing on user comfort and adherence.
- To provide a valuable resource for researchers developing integrated fall detection systems within footwear.
Main Methods:
- Design of a novel ankle-placed fall detection device.
- Identification and selection of relevant activities for dataset collection, including falls, falling risks, and ADLs.
- Dataset collection from 21 users performing the selected activities.
- Evaluation of the collected dataset's quality.
- Implementation of a preliminary Deep Learning classifier to assess system feasibility.
Main Results:
- A new, balanced dataset for fall detection and ADLs was successfully collected using an ankle-worn device.
- The dataset includes data from 21 participants engaged in various activities.
- Initial evaluation indicates the dataset's suitability for training fall detection models.
- A basic Deep Learning model demonstrated the feasibility of using this data for fall detection.
Conclusions:
- Ankle-placed devices offer a promising alternative for elderly fall detection, improving comfort and compliance.
- The newly collected dataset is a valuable resource for advancing research in wearable fall detection technology.
- The study validates the feasibility of using ankle-based sensor data and Deep Learning for effective fall detection systems.
