Ciência de situação na IOT: uma arquitetura explorando processamento híbrido de contexto

Abstract

The Internet of Things (IoT) has influenced the development of computational systems, enabling a more proactive interaction with users, expanding features as mobility and availability. In this scenario increases the demand for applications that can recognize the user’s context and can provide situation based services. The identification of situations is a research challenge for applications in the IoT, given the complexity of the relationships that must be established and processed. Data from several sources can be used for contextual processing, which can occur in multiple steps and involve different techniques based on specification, learning or hybrid models in which both are combined. The use of middleware is highlighted in the literature as a way of treating heterogeneous devices in the IoT and also to make more transparent the contextual processing for the applications. However, for a middleware be independent of the application domain is necessary to add in its architecture layers that allow the management of software components used in contextual processing and still allow the composition of different contextual processing flows. Considering this demand, the aim of this work is to design an architecture for management and composition of hybrid contextual processing flows to provide situation awareness for IoT applications. One of the premises of the architecture is its integration with the middleware EXEHDA (Execution Environment for Highly Distributed Applications). The evaluation of the architecture was done through two scenarios of use, one in the area of cardiac rehabilitation and another in the area of management of hospital environments. The results obtained were promising, pointing to the continuity of the research.