Research On Information Correlation Modeling Technology For Social Internet Of Things

In the Internet of Things, the next generation of Internet can effectively promote the harmounious interaction between people, socities and smart objects. The research on interaction between smart objects facilitates the conformation and the development of Social Internet of things, mainly related to real-world perception, data transmission, trust management, data sharing and service management etc. Similar to the SNS(Social networking services) of human beings, Social Internet of things(SIoT) provides the definition of social relationship and the trustworthiness between smart objects, to ensure the mutual understanding of the shared information. While the relationship management allows objects to choose relationships with others, the service discovery can take advantage of the social relationships of objects to find some individuals can provide the required services. Then the service composition activates the services for ultimately providing applications. The self-perception and social relationships of objects in SIoT are the resources and basis on which we establish an information correlation model. The overall network architecture can compensate for the lack of a single smart object and provide more opportunities or constraints, making a higher integrity of sensored information data. There is also some interference between different smart objects, while the information correlation model for SIoT can be helpful in understanding how social relationship patterns influence the intelligent individuals, as well as the extent of the impacts. These will bring about long-term significance on the overall analysis of sensored information and data.On the basic of the above research background of SIoT and needs for the establishment of the information correlation model, this thesis presents a information correlation model of SIoT, and introduces the detailed function description of the physical infrastructure layer, dividsion processing layer, extraction-transmission layer, information processing layer and application layer. Our information correlation model utilizes the relevant theory of social networks to analyze the social relationships between objects, as well as the location-aware information data. For the established network of sensing devices, we make use of the sensing community division method based on multiobjective clustering to detect the communities, and offer the identities and specific descriptions of the sensing community division algorithm. In addition, for the location-aware data transmitted to the information processing layer after being processed by PML(Physical Markup Language) normalization, we propose a location information analysis method based on collaborative filtering in this thesis. By analyzing the mass location data, we obtain a matrix of location correlation that is the basis for providing LBS(Location Based Service) in the application layer, such as the service recommendation. Then we describe the related identities and the detailed algorithm description of location correlation computation. This thesis also provides the simulation results of these algorithms which well demonstrate the effectiveness and efficiency of them.