Design And Key Technologies Research Of Home-based Care For The Aged System

For the Chinese,the majority of the elderly people will choose the home for their first choice for the pension,seek for the care from the member of the family is the main way to fulfill the pension.However,the size of the family is smaller and smaller day by day.The family structure changed,the number of the empty nest elderly person who live in the urban and rural is increasing,for these reason,the functional of the home-based cared for the elderly people get more and more weak.because the children are busy with their work,they have no time to take care of their elderly people.Besides they have no such related skills either.And that makes the elderly people choose the home for their first choice for the pension became very hard.Nowadays the mobile devices develop very fast.The sensor accuracy on the intelligent equipment is higher and higher.To solve the problems above,enhance interaction between the elder people and their children,this thesis presents a brand-new way to fulfill the dream of the old man living in home for the pension: A mobile APP.This APP contents the old people's basic data detection,geo data detection,action trajectory detection,etc.Besides the APP also contents the fall detection function which based on the acceleration sensor and the gyroscope.The fall detection can get the geo data for the first time when the old people fall down and send a message to their children so that they can handle the serious situation.In this thesis,through the analysis and classification of people's daily activities,we summarize the human behavior into two categories: Activities of Daily Life(ADL)and fall.To solve the problem is to determine the human body ADL and fall activity.The fall detection algorithm in this thesis is designed for real-time fall detection algorithm,and compare the experimental data according to the set threshold value according to the data of acceleration sensors for acceleration data and gyro attitude angle.And at the meantime,in order to judge more accurately,this thesis selects a number of eigenvalues to judge,and puts forward a new method of extracting FFT feature MFFT,which is about 10% higher than the standard FFT coefficient.After a series of experiments,the recognition rate of the fall detection algorithm is 86%.