Intelligent home control system based on BP neural network speech recognition

摘要

With the rapid progress andwide application ofcutting-edge technologies such as automatic controltechnology and wireless communication technology inrecent years, smart home, as an effective combination ofthese cutting-edge technologies and daily life, has receivedmore and more attention and Research has been greatlydeveloped. In order to solve the shortcomings of theexisting home control need to download various cumbersomeAPP, the WeChat applet is introduced into the smarthome system, and the intelligent voice technology isplanned to be used in the smart home control to realize thevoice control, improve the user experience and make thehome Life gets smarter. This time, the deep learning technologywas used in the development process of the smarthome control system, and according to the historicalinformation of the home, combined with the learningability of the recurrent neural network for time series data,through the mining, analysis and learning of historicaldata, the construction based on specific The user’s uniquecomputing model can seamlessly connect the capabilitiesof the voice cloud platform with the capabilities of theInternet of Things cloud platform through the Web server,making it possible to quickly access the voice recognitioncapabilities to smart homes. The entire system has stableconnectivity, easy deployment and low cost. The mainfunctions are deployed in the cloud and extended. After 10rounds of iterative training of the Attention-GRU model inthis paper, its prediction accuracy can quickly rise to about97, and finally stabilize at about 98.2, and the lightingprediction accuracy can reach 95 or higher.