A home monitoring system for elderly people based on MEMS sensors and wireless networks

In order to realize home monitoring of elderly people and protect their personal privacy at the same time during their daily life, a novel ambulatory monitoring system was designed based on MEMS sensors that can continuously record body movements and detect body postures (walking, standing, sitting, falling and lying). The system alarm was triggered once accident “falling” happens, which ensure the monitored person can get timely and effective help. The monitoring system consists of three sensing units. Acquired data were transmitted to the computer wirelessly. A new signal processing algorithm was adopted to detect movement patterns related to the basic body postures and different postural transitions during daily physical activity. A fuzzy classifier was further designed to classify the postures. The experiment results show that the sensitivity and specificity of classification of body postures were: 99% and 99.5% for walking, 95.1% and 96.5% for standing, 99.3% and 99.7% for sitting, 95.3% and 97.1% for falling, 100% and 100% for lying.