Fall-detection performance varied by sensor type and analysis method
The review ranked wearable sensors lowest and deep learning methods highest on detection measures, without establishing benefits in everyday care.
Based on the published abstract. The full paper may contain additional methods, results and limitations.
The 30-second takeaway
This review compared how well fall-detection technologies recognized falls across published studies. Wearable sensors ranked lowest in the reported analysis, while deep learning methods ranked highest across the detection measures. These technical comparisons do not establish fewer falls or injuries, and the abstract cannot identify a best device for an individual.
The abstract gives no numerical performance estimates, uncertainty ranges, or details of testing settings. An abstract-based comparison cannot establish how large the differences were or how systems would perform in a particular home.
What researchers found
Fall-detection accuracy with deep learning
Versus machine learning, threshold methods, and other approaches across reviewed studies; testing timeframe was not reported.
What performance meant
The review extracted measures of fall identification and the time required to train and test systems. These are technical outcomes; they do not, by themselves, demonstrate fewer injuries or better day-to-day safety for older people.
How comparisons were grouped
The review classified sensors as wearable, non-wearable, or hybrid. It also grouped analytical approaches into deep learning, machine learning, threshold methods, and other methods. These classifications distinguish the sensing system from its method of analyzing data.
The original publication
Fall Detection in Elderly People: A Systematic Review of Ambient Assisted Living and Smart Home-Related Technology Performance.
Gorce P, Jacquier-Bret J
Sensors (Basel) · 2025
- PubMed ID
- 41228764
- Record checked
AI-assisted research and writing. This explains one selected publication; it is not a complete review of everything known. Our approach.
One more question, understood.
Keep track of the research you’ve explored.