Data and Decision Intelligence for Internet of Things: Putting Human in the Loop

Abstract

The requirement and expanding needs of bridging the physical and cyber world promote the emergence of a new global internet-based information architecture, the Internet of Things. The main challenge in Internet of Things is how to make fast and accurate decisions based on multi-modal sensing data. In this paper, we put human into the process and formulate a closed-loop computing paradigm for IoT data and decision intelligence. We first propose a human-in-the-loop reference model, which extends the traditional cyber-physical interaction extended into a closed-loop process based on cyber, physical and human three factors. We defined the three-factor decision paradigm in the model. On the basis of the reference model, we analyze the feature requirement in the data intelligent process and summarize it as three key aspects: semantic, interactive and iterative. Further, we analyze several major challenges facing IoT data and decision intelligence from the perspective of data characteristics. We select and review the recent progress in three typical application domains and analyzed whether the current research has solved the target issues. According to the results of the review and analysis, the paper finally summarizes and addresses some key research directions in the future.