ACAIOT: A Framework for Adaptable Context-Aware IoT applications

Abstract

Recently, the Internet of Things (IoT) and context-aware IoT applications are involved in various domains such as healthcare, traffic, and smart homes. The main challenge, while developing context-aware IoT applications, is managing the massive amounts of data and events to get relevant context information. This paper proposes a semantic-based and domain-independent framework called ACAIOT. It hides the details of context management from the high-level services. It proposes a programming abstraction for adapting services' execution to various domains easily. We demonstrate the features and effectiveness of ACAIOT by its execution on a real smart home dataset. We evaluated ACAIOT against previous work in developing a set of services such as monitoring, prediction, and notification. Results show that ACAIOT achieves an average F1-score of 0.82, which is comparable with the current state-of-the-art methods. Moreover, ACAIOT is able to incorporate important compound activities that cannot be handled by previous work.