Study On Sleep Health Based On Intelligent Mattress

Human beings spend about 1/3 of their time in sleep,the quality of sleep affects the efficiency of work and study.Relevant equipment is urgently needed to monitor the health of the night sleep.Polysomnograph is used to monitor the quality of sleep in hospitals,but it needs to be exposed to the human body and is expensive,so it can not be widely used.The intelligent mattress has the advantages of no need to contact with the human body and the high accuracy of monitoring.It can be widely used in the nursing home and the ordinary family.Body movement affects people's sleep quality,the recognition of body motion events is of great significance for the sleep monitoring system.In this paper,three kinds of body motion recognition algorithms are proposed based on different application scenarios.The rule algorithm makes full use of the hardware,the design is simple.The algorithm based on threshold is mature and robust.In this thesis,the problem of body motion recognition is transformed into a classification problem in machine learning,and the SVM algorithm is used to identify the body movement.The rate of respiration is the basic index for evaluating the quality of sleep.There are many challenges in calculating the rate of respiration,such as body movement,different individuals and the hardware.In order to solve the above problem,a method of calculating respiration rate based on peak interval is proposed in this thesis.The algorithm has the advantages of low consumption of hardware resources and high calculation accuracy.Apnea is a sleep disease that is harmful to people's health.It is often at night and is difficult to find.Research on apnea recognition algorithm based on smart mattress is less.In this paper,online learning algorithm FTRL is introduced into apnea recognition,and combined with practical application scenarios,it is improved and optimized.The algorithm uses the time and frequency information of the signal to construct the feature to seek the apnea pattern hidden in the signal.The algorithm of two modules of respiration rate and body motion recognition has been embedded into the hardware system,and the apnea recognition algorithm has also been applied.These sleep indicators can make mattress users know about their sleep health,if they find sleep health problems,they can be solved in time.