Fall Detection Research And Design Based On Muti-sensor Fusion

China has entered an aging society.The rapid development of economy and society has led to a lot of Empty Nesters.Pension problem has become increasingly intensified.Therefore,the health and safety of the Empty Nesters has attracted more and more attention of the society.The social atmosphere that people are afraid of helping the old who are falling down,makes the elderly can not be treated in a timely manner and missed the best time to treat.So it is very important to let the family members know when the old people fell down.In the daily care of elderly people,the falling of old people caused by blood pressure and other chronic diseases accounts 23%.Therefore,heart rate and blood pressure is also an important physiological index to assess the health status of the elderly.This paper intends to investigate a monitoring system for monitoring the fall of the elderly.Through the multi sensors fusion,acquisition of human body posture and by attitude decision threshold of physiological parameters acquisition,extraction features of these parameters values,to achieve high accuracy real-time fall detection system,realizes the elderly fall monitoring system design.Fall monitoring system includes three important parts: hardware platform design(includes hardware circuit and software algorithm),algorithm and program control section.In the aspect of hardware design,the design of sensor module,signal processing circuit and data acquisition system are designed.ARM microprocessor is the core of the whole hardware circuit,its main functions include: multi sensor data acquisition,remote data transmission,geographic location,program control,etc.Through the multi sensors fusion,acquisition of human body posture and by attitude decision threshold of physiological parameters acquisition,extraction features of these parameters values,to achieve high accuracy real-time fall detection system,realizes the elderly fall monitoring system design.Software system is based on Android system.It improves the remote monitoring function.When the elderly fall through remote communication,the family members can get the warm from the telephoned and exactly know where the old people is.People can also use surveillance video to check whether the old fell down.Verified by experiment,fall monitoring system based on multi sensor fusion presented in this paper can improve the accuracy of the rapid fall discrimination.In this paper,according to the different human body to do a threshold processing,research and analysis of the slow fall,put forward a slow fall recognition model,and the simulation verifies the slow fall.