Multimodal data fusion model for smart city

With the rapid growth of cloud computing and big data, in addition to the urgent demand for city development, smart city construction has become one of the hot topics of domestic and international computer science researches. With the increasing number of closed-circuit televisions and sensor devices in urban city, types of data that people can obtain in city increase as well. The city data has multimodal properties like time dependent, heterogeneous, multi-source and high-dimension. How to make the multimodal city data connected, related to each other, and interconnected to each other, and how to mine better and various information for city construction become the key in this area. In this paper, we propose a multimodal data fusion model for smart city: the multimodal connecting growing fusion (MICROS) model. We present our model in three directions. First, targeting at multimodal data, we describe four features: multisource, heterogeneous, time-dependent and high-dimension. Second, we construct the three-layer fundamental model structure for multimodal fusion from bottom to top, including service-information description model, meta-data model and data-connection model. Finally, based on this three-layer fundamental model, we propose a multimodal data fusion model suitable for smart city construction.