EXEHDA-HM: uma abordagem composicional para processamento de informações contextuais em modelos híbridos de contexto

Abstract

With the dissemination of computer networks, and the increased usage of portable devices of high computing power and broad spectrum of use, began the third era in the evolutionary chain of modern computing, ubiquitous computing. A key feature of ubiquitous computing is the context awareness, which can be divided into three stages, being they, acquisition, modeling and processing. The synergy between these three stages boosts the use of context by the applications, among other aspects, the employment of context provides a better user interaction with the application. It is noteworthy that in the context modeling stage, various approaches can be used, and has not yet been found a model that can be applied of manner equally satisfactorily in different application scenarios. Thus, have arisen the approaches based on integration models, hybrid modeling, which combine two or more modeling approaches in order to take advantage of the strengths of each model. The use of hybrid context models brings new challenges for the applications context-aware as they become necessary mechanisms to facilitate access to this information, ways to maintain the persistence of the same, as well as strategies to process the acquired contexts. Thinking about the challenges faced by the use of hybrid context models, this work presents a compositional approach to processing contextual information in hybrid context models, named the EXEHDA-HM, where is provided the use of one repository based on three models: relational; non-relational; and triplestore. Moreover, the proposed approach has as highlights the compositional processing strategy based on rules, which enables an application to combine the contextual information stored in the different models in the same rule, thus contributing to a better identification of the situations of interest.