Design Of Intelligent Home Control System Based On Deep Learning

With the great progress and wide application of advanced technologies such as automatic control technology and wireless communication technology in recent years,smart home,as an effective combination of these advanced technologies and daily life,has received more and more attention and research,and has been greatly developed.Although the current smart home system adopts some timing control and condition control methods,which is different from the previous electrical control forms,its intelligent degree is still very low,far from reaching the goal of autonomous learning and automatic control that people expect.In recent years,the research on artificial intelligence and deep learning is increasingly in-depth.In this case,combining it with smart home provides the possibility for real intelligent home life,and also provides a new direction and rare opportunity for the further development of smart home.Under this background,this paper takes the intelligent home system control as the main line,and carries out the design and research of the intelligent home system.Firstly,the development status of smart home is analyzed,and the research background and significance of the topic are introduced.On this basis,the research plan of smart home system combined with artificial intelligence technology is determined,and the feasible overall design scheme is given.Second,in order to solve the problem that the existing home control needs to download all kinds of tedious apps,wechat applet is introduced into the smart home system,and a device control scheme using wechat applet as the control terminal is designed,the overall structure adopts the form of small program client-server-home equipment,and combined with a variety of wireless communication such as IR,RF,EnOcean and WiFi,realize remote control of household appliances and improved the convenience of equipment control.Furthermore,studies the neural network model based on gating recurrent neural network(GRU)and attention mechanism and its implementation in Tensorflow.With the technology of Tensorflow serving and Docker as the core,the deep learning system deployment architecture was designed and realized,and the complete process of home intelligence prediction model from construction,training,deployment to service was realized.Finally,in order to demonstrate the function of this paper,the remote control function of small program and the intelligent predictive control function of the system are tested.The test results show that the functional modules designed in this paper can meet the system design requirements,and the smart home control system as a whole meets the requirements.It has a certain theoretical and practical reference value.