Abstract. More and more home devices are equipped with advanced computational capabilities to improve the user satisfaction (e.g., programmable heating system, Internet TV). Although these devices exhibit communication capabilities, their integration into a larger home monitoring system remains a challenging task, partly due to the strong heterogeneity of technologies and protocols. In this paper, we therefore propose to reconsider the architecture of home monitoring systems by focusing on data and events that are produced and triggered by home devices. In particular, our middleware platform, named DIGIHOME, applies i) the REST (REpresentational State Transfer) architectural style to leverage on the integration of multi-scale systems-of-systems (from Wireless Sensor Networks to the Internet) and ii) a CEP (Complex Event Processing) engine to collect information from heterogeneous sources and detect application-specific situations. The benefits of the DIGIHOME platform are demonstrated on smart home scenarios covering home automation, emergency detection, and energy saving situations.

Abstract-Pervasive applications are characterized by variations in their context of execution. Their correct behavior requires continuous adaptations, accordingly to changes observed in their environment. Some of the existing approaches tackle this problem by adding stabilization mechanisms on the decision making layer. In most cases it remains costly for applications to trigger decision making procedures, specially when application changes frequency is potentially high. We believe that, to provide more flexible and efficient context-aware applications, stabilization issues should be addressed separately from decision making issues. In this paper, we present a flexible stabilization approach as an elegant solution to that problem.