Research And Implementation On Context-aware Architecture For Pervasive Computing

As information technology continues to evolve, such as the development of smart embedded devices and wireless communication technology, advances in distributed computing technology, a new computing model named"ubiquitous computing"have come into being. Pervasive computing emphasizes the integration with the computing environment. People can interact with surrounding environment more naturally using various computing devices. So context awareness is the fundamental feature of pervasive computing.While applications for pervasive computing are complex and diverse, each one must have the feature. Therefore adaptive and extensive context-aware architecture have become a hot research topic. First, we use ontology to model the context of computing environment and adopt"layer pattern"to construct the ontology model for high generality. The upper layer of this model consists of the general concepts of applied domain in pervasive computing. The lower layer contains the specific concepts of concrete applications of pervasive computing.On the basis of context modeling, the paper addresses an adaptive and extensive context-aware architecture for pervasive computing. It provides efficient support for acquiring, representing, processing and using the context model. Then the paper describes the detailed function and design of every module consists in the architecture and the interaction between each module. At the same time the scalability and configurability of this architecture is adequately considered.After the architecture construction is completed, this paper gives an implementation framework based on OSGi and R-OSGi. In the framework we analyze the implementation mechanism and detail of every module to construct"smart home", which is a typical pervasive computing application. This application demonstrates the feasibility of the architecture.Device is the supporter for context information acquiring and context-aware behavior in continuously running pervasive computing system. Since device renewal is very frequent in the system, it is necessary to research how to minimize the cost of renewal to popularize the pervasive computing application. Based on the previous implementation framework, we propose a dynamic device updating mechanism that does not need the involvement of system developer to solve the heterogeneous device updating problem. With this mechanism, the context aware architecture will be more general and highly utilizable.