Research On Some Key Techniques Of Context-aware Computing Based On Pervasive Computing

The rapid progress of society and continuous development of computing technology need corresponding computing paradigm. Currently, with the computing resources, network connectivity and the interaction between man and computer service being ubiquitous, the information space where is comprised of computers has been integrated with physical space where human lives and works. So, severe challenge is posed on the traditional computer-centered computing paradigm.The emergence of pervasive computing just meets the needs of this application, and it supports users to transparantly get personized information service at anytime and anywhere. It is because the computing paradigm caters for the trend of technology in the future that this research field has a great potential.The term "Transparent" in pervasive computing not only refers to the visibility of information-aware between users and systems, but also refers to the invisibility of information interaction and the fact that the interacting process is not being sensed by users. The context-aware computing, which provides users with implicit interactive services by perceiving, decision-making, reasoning about contextual information, is an important means to achieve this interaction. Therefore, the context-aware computing is the basis for human-computer interaction, and is one of the core techniques for pervasive computing. Aiming at the demand of context-aware applications under pervasive environments, this dissertation studies a number of key techniques for context-aware computing, and its main work is described as follows:(1) Firstly, the context aware computing is analysed for the current situation and the related techniques for it is mainly discussed in this paper. These key techniques include mainly the contextual definition acquisition,modeling,evolution and communications environment, and so on.(2) This dissertation analyses the existing method of context information representation and modeling, puts forward an approach to describing context information through XML, and modeling context information by using RDF. This approach not only is able to achieve the unity of context information representation, but also is able to achieve the necessary expansion for context information modeling. So a good means is provided to represent and model many different kinds of context information with dynamicity.(3) This dissertation compares the existing techniques for the context information filtering, integration and inference. Firstly, the filter for context information uses Bayesian-based method which can effectively remove redundant context information; Second, this dissertation uses the rules-based fusion strategies to fuses the context of information which is stored in the semantic knowledge base. The method is not only valid for the context information to avoid the integration process of a large number of calculations, and the rules and semantics are more easily understood and achieved by users; Finally, this dissertation uses the rule-based reasoning method to reason about the context information. It has the some advantages, such as intuition, reasoning clearly, effective expression of heuristic knowledge and strong modules, etc., and can effectively solve poor reasoning and uncertainty issues for the reasoning about context information. (4) Finally, this dissertation describes the application of a number of key techniques of context-aware computing in smart parkings. We apply the context-aware techniques to the intelligent parking management, and the experimental results in this dissertation show that our approach is feasible.