Smartphone-Based Fall Detection Monitoring System Design Among The Elderly

In recent years, with the rising proportion of the elderly population, more and more "empty nesters" appear. The problem of elderly is concerned, especially the elderly health problem has caused a common concern of the whole society. Falling is the primary harmful factor to the elderly. The consequences of accidents caused by falling not only cause great harm to the elderly, but also make the family bearing irreparable pain and heavy economic pressure.Smartphone-based elderly fall detection monitoring system mean to using the internal inertial sensors of smartphone monitoring elderly’s daily life stance, and detecting falling behavior accurately, and then using smartphone’s location module and communication module alarming others. In this way, it can improve the ambulance speed of falling accident.The research contents of this paper include the following aspects:(1) Paper have done the research of the Android smartphone’s build-in inertial sensor including accelerometer and gyroscope, and collecting original attitude data of falling action and daily behavior using the Android SensorManager interface. And paper then design median filter and high-pass filter getting noiseless linear acceleration data and angular velocity data.(2) Paper propose a fall detection algorithm based on following eigenvalues: SVM(Signal Magnitude Vector), SMA(Signal Magnitude Area), CST(Continuous Static Time). The first stage is the decision tree fall detection discrimination law was used to Judge human behaviors using Instantaneous data, and the second stage is curve analysis of historical data to judge.(3) Paper have done the research of the Android smartphone’s build-in GPS sensor, using the GPS sensor locating the position data, and then sending alarming message to the preconfigured associated equipment by 3G / 4G mobile networks. At the same time Smartphone’s speaker will make noise to remind people to assist treatment.(4) Paper has made comprehensive Android application,which has beautiful, simple and easy interface, and has following feature: Real-time sensor data monitoring, background fall detection, integrated positioning, SMS alarming and sound alarming.Testing showed that the decision tree fall detection discrimination law can achieve 97.3% comprehensive correct rate, and other features reached the system’s requirement.