Embedded intelligent pension system based on deep learning

In this paper, according to the living habits of domestic ordinary families, room layout and other specific circumstances, put forward four basic points of intelligent family pension plan to help families to achieve intelligent home care. First of all, the system not only provides rich ZigBee intelligent devices for the elderly to meet their needs in daily life, but also integrates emergency keys and Kinect devices into the system, so that when the elderly have accidents, they can send alarm information to the outside world to ensure their safety. Secondly, an intelligent pension system based on health data mining is proposed. Firstly, the heartbeat signal and respiratory signal of the human body are collected by the non-contact vital sign sensor, and the weight signal of the human body is collected by the pressure sensor. The collected physiological parameters of the human body are sent to the remote server based on the wireless module. Then, by comparing the efficiency of classification and clustering methods in heart rate and respiration rate data mining, a health data mining algorithm is implemented based on K-means clustering. Considering the reliability and stability of the system, the upstream and downstream data between the intelligent device and the cloud server during the operation of the system are all governed by the unified communication protocol rules. Finally, users can interact with the system through Web pages, smart speakers and mobile terminals, so as to give full play to the diversified manipulation characteristics and increase the elderly home care intelligent life experience.