| With the development of economy and society,the proportion of the elderly population is gradually increasing,and intelligent elderly care will become a social development trend.How to use modern information and communication technology,such as IoT,big data,artificial intelligence,distributed deployment,etc.to achieve intelligent management of elderly care services is a hot research issue in smart elderly care.Based on the infrastructure of the existing Smart-Elderly-Care service platform,this thesis focuses on the commonly used smart terminal equipment,and designs and implements the development plan of the smart pension mobile APP from three aspects:client,background service,and edge.The plan includes extracting public components in a component-based development method,building a mobile client involving multiple user roles;proposing a design and implementation scheme of the Database Module and Interface of Function Module;implementing an elderly fall detection service using edge computing and computer vision.which optimized and improved the target detection model from the anchor setting mechanism and the positive and negative samples determination method,and improved the accuracy of elderly identification.The main work of this thesis is as follows:(1)Based on the actual needs of elderly care life,facing the multiuser roles of the elderly,children and the operators the Smart-ElderlyCare service platform,this thesis proposes a design and implementation scheme of the client APP in the smart elderly care service platform.It adopts the component development strategy,extracts the public components,and integrates different general-purpose APPs.The layered design of functional components at a high level can realize functional combination,reduced the redundant work of developing multiple clients,and improved the reusability and expansion of functional components.(2)A design and implementation scheme of the background functional module framework of the mobile APP in Smart-Elderly-Care service platform is proposed,which is constructed using the Spring Cloud framework as a whole.On each functional module,the Spring Boot-based microservice architecture is used,and the structure is hierarchically designed from top to bottom to provide external callable interfaces.a design and implementation scheme of the background service data storage of the mobile APP in the Smart-Elderly-Care service platform is proposed.It realizes functional modules of the platform such as basic information management,abnormal alarm,service management,order management,and message synchronization.(3)A design and implementation scheme for elderly fall detection using edge computing and computer vision is proposed.It considers the server computing pressure of the smart pension cloud platform under multi-user conditions,based on the target detection model,combined with the change of human posture to determine the fall state,after weighing the calculation accuracy and time delay,the SSD＿MobileNet＿v2 network is selected as the target detection network,perform data enhancement processing on the collected data set,and after the trained network is lightweight,packaged and deployed to the edge side in the form of docker.According to the characteristics of the singletype target detection scene of the Smart-Elderly-Care service platform,the inherent setting method of the anchor in the SSD algorithm is improved,and clustering algorithm is used to analysis the central value of the aspect ratio of the data set.ratio of the data set.According to the clustering results Re-set the anchor aspect ratio in the SSD algorithm to improve the recognition accuracy.Aiming at the problem that the positive and negative samples of the anchor in the current SSD algorithm ignores the characteristics of the data set itself,a method of adaptively determining the positive and negative samples in the SSD＿MobileNet＿v2 network is proposed to improve the fall detection function in the smart elderly care service platform.Theoretical analysis and experimental results show that the accuracy of fall recognition for the elderly is improved. |
