Research On Location-aware Service In Pervasive Computing

Pervasive computing, integrating of computer, communication and digital media technology, makes it possible to merge the physical world we are living in and the virtual world in the information space. The basic idea of pervasive computing is to embed computer into the living environment or tool of human, so as to make the computer invisible from users' sight and the users can only focus on their task instead of the computer. Location is an essential part of contextual information, which has turned out to be useful in many applications, particularly in determining position, navigation, routing, tracking, logistics, and monitoring of pervasive computing devices. For example, turning on the light near the user automatically, or transferring the phone to the nearest user and so on. All of these applications should be able to determine the user's position and the relative position to other devices. Consequently location-aware computing is one of the key but challenging areas of pervasive computing. In this dissertation, we focus on location-aware service in pervasive computing. Our approaches and contributions are given below.All of existing location-aware systems, positioning algorithms and positioning technologies are analyzed. The positioning system based on Ultra-wideband (UWB) is proposed. We deduce Cramer-Rao lower bound (CRLB) of UWB distance estimation, and estimate distances using maximum likelihood estimator (MLE). All results of our work are simulated by MATLAB. These results demonstrate that the positioning system based on UWB successfully reaches centimeter scale in precision, and MLE is a simple and effective method for location-aware positioning. After analyzing the existing location models, we construct a novel location-aware model based on ontology, which can integrate heterogeneous location-aware positioning systems into pervasive computing environment. The Location-aware Model Ontology (LMO) is also proposed in this dissertation. By using the Web Ontology Language (OWL), LMO describes location-aware context vocabularies including space, time, user profiles, and policies for security and privacy. So the location-aware information can be defined exactly, expressed clearly, and organized effectively in a uniform semantic mode. Context knowledge can be shared and re-used in security.The reasoning between the location-aware context knowledge and its properties is carried through using ontology reasoner. By using FaCT++ (Fast Classification of Terminologies) reasoner, the classification hierarchy for the LMO classes is calculated, and inconsistencies of the LMO classes are checked. The relationships among the properties are defined through the rules coded in OWL and SWRL (Semantic Web Rule Language). The deeper reasoning of properties' relationships is performed by Jess Rule Engine.A context access control model (CACM) based on UCON_ABC (Usage CONtrol) is presented. CACM not only focuses on authorization, but also introduces obligations, and conditions. Three integrated sub-models are extended flexibly, including pre-authentication & pre-condition model, active-provision & passive-provision model, and pre-authentication & pre-obligation model. The security, usefulness and flexibility of CACM are demonstrated by three typical location-aware application scenarios.A dynamic and semantic security policy management mechanism for location-aware computing is presented. The mechanism extends the existing policy model by importing context provider and ontology service. The context changes of various entities in location-aware computing environment can be obtained by context provider. According to the context changes, ontology service updates the information of policy service and event service automatically. Then policy service adjusts the action on the managed targets by policy enforcement agent dynamically. The mechanism solves the contradiction between information sharing and security privacy in the location-aware computing environment.