The Research And Design Of Smart Home System Based On LonWorks

With the development of information technology, the intelligent control technology is expanding day by day. Smart home is already a necessary direction.The integration of Internet, Intranet and Infranet, which will upgrade the intelligent level of every home and uptown, and will influence manner of production, life and work of people and promote society developing.Smart Home is still in initial phase in our nation. Its developed technology need to be introduced into our country, and learned by us. Therefore, it has profound theory significance and applicant value that researched a project and architecture based on case of smart district in our country.The first, it introduces concept and develop status of Smart Home. It mostly present strongpoint of LonWorks fieldbus technology, neuron chip, LonTalk protocol, Neuron C programming language and LonWorks developing tools after the concept and use of fieldbus are introduced.According to requirement of intelligent home, it expounds the way of designing the Smart Home network controller with many uses, which are gathering data, control devices and communication. It can reduce workload in creating field bus network, and easy to maintained. A new structure that connect with Web server, SQL database, management system, and Lonworks fieldbus network. Moreover, the system structure of home network of intelligent uptown is bring forward, and expatiate the process of designing muti-input/output Smart Home Controller. The management subsystem in host is part of the Smart Home system. It cans automatic reading data from water meter, ammeter, and gas meter, and responds to alarm signal. Another part is remote control subsystem, which can organize or cancel defence, view alarm record, and query charge of water meter, ammeter, and gas meter.Lastly, it discuss RBF(radial basis function) neural network appliance in intelligent detect alarm. A RBF(radial basis function) neural network be devised that use to intelligent infrared alarm barrier. Moreover, K-means and gradient descent learn algorithms be used in network learning process. The result is that, by Matlab simulation experiments of using K-means and gradient descent in infrared hurdle, RBF neural network learning algorithm based on K-means have high precision and rapid execution speed, and it can meet actual needs.