Smart Wearables with Sensor Fusion for Fall Detection in Firefighting
<- Publications
Smart Wearables with Sensor Fusion for Fall Detection in Firefighting
Xiaoqing Chai,
Renjie Wu,
Matthew Pike,
Hangchao Jin,
Wan-Young Chung,
Boon-Giin Lee
Sensors | 2021
| View on Publisher's Website
Abstract
During the past decade, falling has been one of the top three causes of death amongst firefighters in China. While there are many studies on fall-detection systems (FDSs), most rely on a single motion sensor and have not fully explored sensor placement and positioning effects. Existing solutions mainly target elderly populations rather than high-risk professions like firefighting. This study proposes a smart wearable FDS for firefighter fall detection by integrating motion sensors into personal protective clothing at the chest, elbows, wrists, thighs, and ankles. A multisensory recurrent neural network model is used to detect falls, and various sensor placement configurations were tested for accuracy. The results show that a fusion of sensors across all five body parts achieved 94.10% accuracy, 92.25% sensitivity, and 94.59% specificity. The study highlights the potential of wearable sensor fusion to enhance firefighter safety and improve fall-detection performance in challenging environments.