Research On Information Fusion Technology In Internet Of Things For Context Awareness In Shmart Home

With rapid development of microelectronic technique and communicationtechnology, it is shifted from traditional computer-centric system to human-centricsystem, which is called Pervasive Computing. Now more and more people put theirattention to a new concept of ‘Internet of Things’, which is based on the pervasivecomputing. People can use different kinds of terminals, such as telephone, PDA,computers and so on to link to Internet of Things as need as possible. In this way peoplecan know much more surroundings information anywhere and anytime. Home is anecessary place for all of us. So how to make it more comfortable and unrestrained tocater to our requirement is so important. This is a right place in which Internet of Thingscan play a part. Internet of Things can deployed a great quantity of deferent kinds ofsensors in a family to gather much information about changing environments. Moreover,how to dispose of so much changing information, which came from Internet of Thingsdeployed in a family, so as to let them reflect these changes more truly and timely is akey point we should consider. And data fusion is an essential link in this process. Manypersons are studying on principle of data fusion; however, there are few persons deeplymake research on data fusion within a family environment. Because human is the centerof family environment, we not only need to consider how to objectively describesurrounding changes in family, but also take influences of human into account whichraises the difficulty in study. So in this background, this thesis presents a way of datafusion to solve above problems.Firstly, we present a total technique framework about smart home system in which‘smart magic box’ is center server, and represent many devices and sensors which can bedeployed in a smart home, and also introduce how to communicate between devices andhow these devices compose an integrated system. In this part, we present a framework ofmiddleware, which disposes data came from Internet of Things and analyzes the orderdelivered from users or other models in system and also send these orders to relativehardware.Secondly, based on presenting a system framework, a smart home system can befulfilled by accurately apperceiving what is happening in a family. For this purpose, toget data from home is first step. So, in this part, we classify the data which can begathered in a family, and analyze their respective characteristics, then provide the methodof denoising according to these characteristics. Single or multi data resource willdescribe some states in a family. We call these states events. So we need to define theseevents and trigger conditions of event, and show a way of how to match these conditions.At last, as a result of home where smart home system deployed in, we demonstrate how to acquire and update a user preference (e.g. tempreture, light et al.) by self-learningbecause of important influence of human.Lastly, we design and realize a smart home system middleware which is responsiblefor date transform and data fusion. Also we introduce framework of middleware and itsfunction that it has realized. We can configure devices which are deployed in family by it,also set and modify user preference.