Addressing Context Awareness Requirements of IoT Systems: A Model-Driven Approach

Context awareness (CA) plays a crucial role in Internet of Things (IoT) systems, enabling them to dynamically adapt based on environmental, user, and system states. The elicitation, specification, and management of these CA requirements is challenging due to their inherent complexity, dependencies, and potential conflicts. Existing approaches primarily focus on runtime adaptations but lack structured methodologies for early-stage modelling and integration of CA requirements. This paper introduces a model-driven approach that systematically incorporates context awareness into the requirements engineering process. It leverages large language models (LLMs) for automated elicitation and classification of CA requirements, enabling the discovery of new adaptive services that emerge from system context awareness. Our work builds on the UCM4IoT requirements modelling language tailored for IoT systems. We also propose a domain-specific language, CARDML4IoT, for modelling and analysis of dependencies in CA requirements. Our proposed approach is demonstrated with a smart store system application.