Enabling Convenience in IoT-based Smart Homes

Abstract

We present a novel framework for enabling convenience for the residents in IoT-based smart homes. Enabling convenience in smart homes requires the study of three main research tasks: IoT service usage pattern discovery, IoT-based service composition, and conflict detection and resolution. The three tasks are interconnected, i.e., the outcomes produced by the first task serve as knowledge aid for residents to personalize service composition; Conflicts may arise when several service compositions execute simultaneously. In this regard, the third task is needed to detect and resolve conflicts. Specifically, the task of discovering IoT service usage patterns focuses on extracting IoT service relationships and IoT service periodical patterns from human-IoT service interaction history. We propose a new ontology-based hybrid approach combing the strength of the knowledge-driven and data-driven techniques to discover a variety of IoT service relationships. We propose a novel approach to discover periodic composition patterns from IoT service usage history considering the quality of the discovered candidates. As for the task of IoT-based service composition, we propose a novel approach to empower non-technical residents to personalize smart home environments through composing services using the ECA (Event Condition Action) rule paradigm (i.e., The composite service is also referred to as a trigger-action rule or an ECA rule.). As for the task of conflict detection and resolution, it relies on a knowledge graph model for IoT services and a conflict taxonomy. We propose a new approach to accurately detect conflicts and properly resolve conflicts. Experimental results validate the performance and effectiveness as well as the applicability of these proposed approaches.