A Novel Attitude Feature Extraction Method for Multi-IMU Based Fall Detection System
2023 IEEE 3rd International Conference on Power, Electronics and Computer Applications (ICPECA), March 2023
Xiaoqing Chai, Boon-Giin Lee, Matthew Pike, Renjie Wu, Xian Wu. 2023. A Novel Attitude Feature Extraction Method for Multi-IMU Based Fall Detection System. In 2023 IEEE 3rd International Conference on Power, Electronics and Computer Applications (ICPECA). DOI:https://doi.org/10.1109/ICPECA56706.2023.10075997
Xiaoqing Chai and Boon-Giin Lee and Matthew Pike and Renjie Wu and Xian Wu. (2023). A Novel Attitude Feature Extraction Method for Multi-IMU Based Fall Detection System. 2023 IEEE 3rd International Conference on Power, Electronics and Computer Applications (ICPECA). https://doi.org/10.1109/ICPECA56706.2023.10075997
Xiaoqing Chai and Boon-Giin Lee and Matthew Pike and Renjie Wu and Xian Wu. "A Novel Attitude Feature Extraction Method for Multi-IMU Based Fall Detection System." 2023 IEEE 3rd International Conference on Power, Electronics and Computer Applications (ICPECA), 2023. https://doi.org/10.1109/ICPECA56706.2023.10075997
Xiaoqing Chai, Boon-Giin Lee, Matthew Pike, Renjie Wu, Xian Wu. 2023. A Novel Attitude Feature Extraction Method for Multi-IMU Based Fall Detection System. 2023 IEEE 3rd International Conference on Power, Electronics and Computer Applications (ICPECA). doi:10.1109/ICPECA56706.2023.10075997
Xiaoqing Chai and Boon-Giin Lee and Matthew Pike and Renjie Wu and Xian Wu, "A Novel Attitude Feature Extraction Method for Multi-IMU Based Fall Detection System," 2023 IEEE 3rd International Conference on Power, Electronics and Computer Applications (ICPECA), 2023. doi: 10.1109/ICPECA56706.2023.10075997
@inproceedings{icpeca-2023,
title={A Novel Attitude Feature Extraction Method for Multi-IMU Based Fall Detection System},
author={Xiaoqing Chai and Boon-Giin Lee and Matthew Pike and Renjie Wu and Xian Wu},
booktitle={2023 IEEE 3rd International Conference on Power, Electronics and Computer Applications (ICPECA)},
year={2023},
doi={10.1109/ICPECA56706.2023.10075997}
}
Fall detection, Wearable sensors, Inertial measurement unit (IMU), Machine learning, Feature extraction, Attitude and heading reference system (AHRS)
Abstract
Falls are a leading cause of preventable injuries, particularly in high-risk professions like firefighting. Traditional fall detection systems (FDSs) using inertial measurement units (IMUs) rely on feature extraction techniques that increase computational load. This study proposes a novel attitude feature extraction (AFE) method that extracts five key feature sets from 54 raw measurements collected from nine IMUs integrated into firefighter protective clothing. The results show that the AFE method significantly reduces processing time while maintaining or improving fall detection accuracy compared to conventional raw feature extraction (RFE) methods. The proposed approach enables real-time, on-device fall detection, making it suitable for constrained processing architectures in wearable safety systems.