点击次数：

发表时间：2022-10-10

发表刊物：生物医学工程学杂志

所属单位：电子信息与电气工程学部

卷号：30

期号：1

页面范围：95-99

ISSN号：1001-5515

摘要：A falling is a risky incident of safety and health of human. It may
   cause serious injuries, such as bone fracture, and even death. A falling
   detection method based on inclinometer is described. At first, we
   collect angle data recorded by a wearable inclinometer placed at
   subject's waist. The angular data are transmitted to PC through a
   wireless data transmission device. Then, the falling duration is divided
   into three phases: the state of fall, the impact phase, and the posture
   phase. We make threshold-based fall-detection decisions in every phase
   after feature extraction and analysis of the short-time angle data.
   Finally, a robust falling detection result is given by comprehensive
   considerations of the three phases decisions. The experiment results
   proved that the accuracy of our falling detection method was up to
   97.23% without undetected falls.

备注：新增回溯数据