The Design Of Elderly Behavior Identification System Driven By Data

Population aging is an unavoidable problem in many social development processes.China has a large population,regional,urban and rural development imbalance.To make the majority of the elderly can be well maintained,the need for the whole society to work together.The development of services for the elderly,it is necessary to ask developers to stand in the old position of thinking.Therefore,the second chapter of this article has made a detailed analysis and summary of the construction of the wisdom of the old-age community,from the analysis of the needs of the elderly,the basic needs of life,health conservation needs to make norms and determine.On this basis,through the scenario planning method,sort out the system of intelligent pension community model,and determine the model to solve the problem,that is,the object "position" and "situation".Community as a place where the elderly daily living,most areas need to take confidential measures,so this study does not involve video and image technology.For the demand of the location,the third chapter of this paper designed the indoor positioning system for the elderly living in the community,the system uses the fingerprint database location method.In the easy to be ignored fingerprint library selection and the establishment of links,innovative introduction of density peak clustering algorithm,in the process of classification and positioning,the random forest algorithm is used to estimate the location information of the location to be determined,and the performance difference between the method and the previous method is compared by using the two sets of experimental data obtained by autonomous acquisition and network.The location of the effective access,so that other location-based diversification services can be effectively carried out,so in this type of system,positioning system has the primary and basic status.The acquisition of the object situation is a higher level of demand,the elderly ability to action,the ability to respond to the decline in the daily life of the probability of accidents increased significantly.Manmade real-time monitoring,high cost,poor flexibility,do not have the promotion conditions.Smart phones,bracelets built-in rich micro-sensors,various types of sensors to collect data types,the number of large.However,the complex data is almost always not fully utilized,the reason is the lack of effective data mining methods.In recent years,the development of deep learning has opened up a new way to deal with the massive data information mining.The release of various deep learning framework makes the construction of CNN and DNN simple and fast.The application scene becomes rich and varied.The fourth chapter is the use of mobile phone sensor as the data source,the neural network modeling method applied to the elderly fall detection system design.The fall detection system only completes the key part of the situation demand.The daily behavior recognition of the elderly still needs new system participation.To simplify the construction of the new system,the behavior recognition system is built on the indoor positioning system and the fall detection system.In the fifth chapter,the traditional HMM is changed to form an HMM with two observation sequences,which correspond to the time series and the time series of 3D human data respectively.The elderly intelligent monitoring service system is through the JAVA,Web,database and other technology development of online service platform,the main use of the object for the community management personnel,the system integrates the three intelligent service system,and called the local data and Web data As a supplement,the pursuit of feature-rich while maintaining a simple and friendly user interface(UI)and user experience(UE).