A Survey on Data Fusion in Internet of Things: Towards Secure and Privacy-Preserving Fusion

Abstract

Internet of Things (IoT) aims to create a world that enables the interconnection and integration of things in physical world and cyber space. With the involvement of a great number of wireless sensor devices, IoT generates a diversity of datasets that are massive, multi-sourcing, heterogeneous, and sparse. By taking advantage of these data to further improve IoT services and offer intelligent services, data fusion is always employed first to reduce the size and dimension of data, optimize the amount of data traffic and extract useful information from raw data. Although there exist some surveys on IoT data fusion, the literature still lacks comprehensive insight and discussion on it with regard to different IoT application domains by paying special attention to security and privacy. In this paper, we investigate the properties of IoT data, propose a number of IoT data fusion requirements including the ones about security and privacy, classify the IoT applications into several domains and then provide a thorough review on the state-of-the-art of data fusion in main IoT application domains. In particular, we employ the requirements of IoT data fusion as a measure to evaluate and compare the performance of existing data fusion methods. Based on the thorough survey, we summarize open research issues, highlight promising future research directions and specify research challenges.