Design And Realization Of Empty Nest Elderly Guardian System

Nowadays,the degree of population aging is becoming more and more serious.With the continuous development of society,science and technology,the demand for security monitoring of empty nesters is also increasing.The use of embedded systems and Internet of Things（Io T）technology can monitor the life of empty nesters and deal with emergencies in time,so as to guarantee the safety of empty nesters.For such social problems,this paper adopts ARM embedded technology combined with sensors,wireless communication and target detection algorithms to realize a guardian system that integrates environmental monitoring,physiological monitoring and dangerous behavior detection for empty nesters.Firstly,in this paper,STM32F103 is used as the main control chip in the design of the sensor data monitoring part.In the environmental monitoring part,all-in-one sensors are used to collect the data of temperature,humidity,light intensity,CO₂ concentration,PM2.5concentration,noise,combustible gas concentration,etc.,and communicate with the main control chip through the RS485;in the human physiological monitoring part,the data of the empty nesters’daily body temperature,as well as heart rate,frequency of respiration and sleep state at rest are collected.The above data are uploaded to the Io T cloud platform through the wireless communication module,and alarming measures in different ways are carried out through the reminder module.Secondly,in the risk behavior monitoring part,the target detection algorithm is used to discriminate the behavior of empty nesters.Using YOLOv5 target detection algorithm combined with Gaussian mixture model for dataset training,and the trained model is transplanted to the Raspberry Pi device equipped with the ncnn framework.The dangerous behavior of empty nesters is identified through the video capture device and the text messages are sent to family members,if there is any abnormality.Finally,the whole system is tested.The sensor data collection part can monitor the environmental data,and the cloud platform can also receive the data collected by the relevant sensors;the dangerous behavior detection part can distinguish abnormal behaviors and send alarm messages to family members.