Research on smart home environment system with multisensor and three level data fusion

A three level data fusion architecture is proposed to facilitate the real-time detection and the intelligent adjustment of living environment, which is based on fuzzy theory, back propagation (BP) neural network, and Dempster-Shafer (D-S) evidence fusion weight optimization method. The machine learning algorithm is used to judge the living environment with historical data, so that the current environment can make users feel comfortable, and obtain the preferred environmental parameter values of residents. According to the living habits of residents, the transformation of environmental parameters from static to dynamic can be achieved efficiently, and real-time dynamic feedback can be realized. The test results show that the system has a detection accuracy of 99.96% with the fast response speed and good reliability performance, as well as the good real-time and adaptive capabilities. Therefore, the system not only provides convenience for users, but also reduces safety hazards in the home environment.