Research On Family Health Detection Method And Application Based On Digital Twin

Smart home is a research hotspot in the application technology of the Internet of Things.However,traditional smart homes mostly focus on smart home appliances and ignore the residents.In view of the above-mentioned current situation,this paper builds a family digital twin prototype system based on digital twin technology,and conducts research on the key technologies of system realization.The specific contents are as follows:(1)In view of the wearing comfort and data validity of human wearable sensors,a reasonable collection plan is proposed.This article first analyzes the types of data needed in the home digital twin system,and then selects the sensor and its wearing position.The nine-axis inertial sensor and the UWB positioning sensor are fused to form a "smart belt";the inertial sensor is calibrated and Filter operation to complete attitude calculation;realize indoor three-dimensional positioning through the two-way time-of-flight method.Finally,the collection and aggregation software was developed to realize the collection and summary of all sensor data in the family digital twin system.(2)Aiming at the problem of low accuracy of motion recognition with a small number of sensors,SVM is used to achieve high-precision recognition of sitting,standing,walking,running and falling.In this paper,through data analysis,the five actions are regarded as periodic motions,and the cross-validation method is used to find the optimal window,and the split method is used to achieve dimensionality reduction of the feature vector;for the SVM multi-classification problem,the optimized directed acyclic graph is more The hierarchical structure of the classification method and a multi-classification method based on the Huffman tree are proposed;experiments show that comparing the one-to-one multi-classification method between the two methods greatly improves the recognition efficiency while maintaining a higher F1-score.Based on a real family,this paper designs and implements a family digital twin prototype system.Use 3DMax to complete the development of home scene modeling,character modeling and character animation;complete the construction and data driving and synchronization of digital twins in Unity3 D,and export Web GL,combined with the Vue.js framework and Springboot framework to realize the fall of the Web side Alarm,twin monitoring and health data query functions.This system verifies the feasibility of this research method from the application level,and to a certain extent provides support for the exploration of new smart home life models in the future.This paper has 54 pictures,21 tables,and 82 references.