| The Fall is a dangerous situation which the older age groups often encounter,and it will bring serious health problem to the elderly whose physical function hasbecome vestigial, and sometimes it may be fatal. International research on thedetection of Falls, especially the Fall prediction research has just started, and thereare many key issues remain unresolved.The Early Warning System of the Fall is Designed to monitor the activities ofthe consumer in real-time. It could detect the trend when the Fall occurres, and senda alarm signal before the body collide with the ground. The alarm signal can makeairbag and other protective devices play the role of a buffer to reduce the impact ofcollision. The Fall early warning system send the alarm signal before the collision，but general Fall alarm system send it after the collision, and that is the biggestdifference between the two kinds of Alarm System.The goal of this project is to design a typical Fall Early Warning andMonitoring system. This thesis is mainly carried out work in the following areas.First, the Mechanical Analysis Model of ADL and the Fall was created to findout the difference between them; And Simulation model of the Fall based on theSoftware of ADAMS-LifeMod was created to get the acceleration curve;Second, the Experimental system including a Triaxial Digital Accelerometer-ADXL345, a Processor-MSP430F149and some other function modules related isestablished. And all kinds of required experiment was organized and executed tocapture the data related.Third, the algorithm based on the Threshold of SVM(Signal Magnitude Vector)was put forward, and it was simulated through the Matlab for its effect. Thealgorithm could make the whole system enough Time Margin and high Accuracy todetect the Fall.Last,the fanal Product with suitable Hardware System and Software Systemwas finished. It was proved to be qualified in the aspect of Accuracy, Error rateand Time Margin through a series of experiments. |
