Research On Key Technologies Of Intelligent Nursing Home Management System Based On Deep Learning

With the continuous development of social economy and the gradual improvement of the medical security system,people’s living standards have also improved,and the aging population will become an inevitable trend of China’s social development.From the macro perspective,the accelerating aging process has brought great pressure to the country and society.From the micro level,the separation of the elderly with care and the elderly with care has caused a series of contradictions and problems,such as difficulty in getting medical treatment and maintenance,to some extent.At present,the living environment,living conditions and supporting service facilities of many elderly people have great hidden dangers.The lack of relevant laws and policies cannot well meet the needs of the elderly for self-realization.With the rapid aging and the increasingly common "empty nest" phenomenon,the pension problem has evolved into a social problem that urgently needs to be solved,and has gradually become the focus of concern of all sectors of society.Based on the analysis of intelligent equipment control and fall detection of the elderly in intelligent nursing homes and in combination with the relevant technologies of deep learning,this paper proposes a complete set of control research process for the elderly from the outside to the inside.The research content of this paper has important reference value and scientific research value for the application of deep BP neural network in the field of elderly auxiliary control.In this study,data simulation was carried out strictly in accordance with the physical signs and conditions of the elderly.Starting from the practical application and scientific research significance,in-depth research was conducted on the key technologies used in the two aspects of the autonomous control of the intelligent equipment the elderly are exposed to and the accidental fall detection of the elderly.The main research content and innovation points of this paper are as follows:(1)combined with relevant technologies of stack self-encoder and deep BP neural network,select more targeted data processing methods for intelligent equipment on the basis of existing research;(2)according to the unique monitoring data and the status of intelligent devices in intelligent nursing homes,appropriate data extraction methods are used to establish the intelligent device control model.(3)unsupervised pre-training of model hidden layer parameters is carried out by using the autocoder correlation method to obtain more abstract features of training data,which can effectively alleviate the gradient dispersion problem existing in the deep neural network.(4)in the study of indoor fall detection of the elderly,the method of nuclear principal component analysis(npca)was innovatively proposed to conduct dimensionality reduction treatment on the features excavated.(5)wavelet threshold noise reduction pretreatment is carried out on the original signal acquired by the integrated sensor,making the acquired human motion characteristics easier to analyze and study.The experimental results show that the accuracy of the method used in this paper is 99.74% for the intelligent equipment of intelligent nursing homes and 99.21% for the accidental fall detection of indoor elderly people,and the specificity is 99.87%.