Design And Implementation Of Smart Home Energy Saving Based On Context Awareness

In nowadays' society,the arrival of the Internet of Things era causes the rapid improvement and progress of the level of smart home in every household as a whole.In the foreseeable future,different types and more and more smart terminals will eventually spread in every corner of each household,thus creating a more intelligent and convenient living conditions for people.Firstly,the traditional home energy-saving system is difficult to meet people's needs such as providing intelligent and personalized service.With the development of context awareness,sensors,terminal devices and other Internet of Things devices,it brings life and hope to deal with that bottleneck.In order to provide users with adaptive intelligent energy-saving control based on current situation information,this paper proposes a smart home energy-saving system based on context-aware,which uses context-awareness engine to design energy-saving module,and realizes the autonomous linkage control and feedback of home devices.Secondly,with the increase of the number of terminal devices in the home and the diversification of users' business needs,the delay and congestion of network will become more and more prominent.Using cloud computing alone can not meet the future development trend of smart home.In order to solve this problem,this paper introduces edge computing technology(EC,Edge Computing)into the system.Modules such as edge data analysis,intelligent perception and control,and electricity demand forecasting,which are based on edge computing can enhance residents'awareness of energy conservation and guide residents to save energy consciously.In addition,GBDT,XGBoost,LightGBM and other advanced machine learning algorithms in recent years are used in our electricity demand forecasting module.Based on open data-sets about electricity,using weather data and working days,holiday time event and other effective information by feature engineering,then a generalized Imodel is constructed in the cloud.Then the model will be extended to the user's side to continue to promote the growth and optimization of the model to obtain the ultimate user personalized model at local.Finally,in order to achieve the above research objectives,Oour system first completes the deployment and Implementation of hardware facilities and software platform including raspberry client and cloud server.Then the overall architecture of the system is designed and implemented,and all functional modules are integrated to the mobile terminal system by using WebApp technology.At the end of this paper,the interface display and the tests of all application functions in the system are given to prove the availability and effectiveness of our system.