Design And Realization Of Generalized Context Inference Engine Based On GaCAM

Pervasive Computing is a new computing mode, the aim of which is to achieve the integration of information space (consisting of communications and computing) and physical space where people live and work. In this environment, people can access digital service anytime and anywhere. It is very important of context-awareness as the premise and basic of the realizing of pervasive computing environment. The middleware technology has long been the research focus of scholars a home and abroad, because it can shield the hardware, software development and system run-time heterogeneity. It also provides a unified access interface for the upper applications. Context-aware middleware is a significant support platform for information acquisition and processing in pervasive computing environment.However, traditional middleware focused on the middleware transparency, which was really convenient to a lot of upper application development in the past, for they could only focus on function to achieve, without having to care about the distribution of sexual and other issues. Many middleware systems, such as Aura, Gaia, PCOM, ScudWare, are applicable only to specific types of upper layer application. However, traditional middleware systems cannot adapt to the current mobile computing environment. There is a great demand for designing middleware systems that can support new requirements imposed by mobility. So we developed GaCAM:a middleware system to support the efficient development, automatic construction and deployment of context aware applications.Context reasoning as an integral part of a context-aware middleware, this paper designs and implements a generic context inference engine-GCIE. As a subsystem of GaCAM, GCIE successful complete the task of context reasoning in GaCAM. And GCIE can also be used alone as a context reasoning engine. Moreover, GCIE has the following features:1) GCIE itself integrates an ontology inference engine, a logical inference engine, Bayesian network inference;2) Supports the expansion of the context inference engines, which allow users add or delete context inference engine based on system needs or personal preferences;3) Effective unified ontology representation. It makes different ontology reasoning engines share the context ontology, which enables different context inference engines work together and substantially increase their reasoning ability;4) Probabilistic extension to context ontology language. It extends the ontological expression of uncertainty and enables Bayesian networks make effective reasoning in the ontology.