Survey of data fusion in IoT
With the advent of the Internet--and, in particular, the World Wide Web--came a massive increase in the amount of data we as a society create on a daily basis. The more recent proliferation of mobile devices and online social networking has increased the amount of user-generated content to an enormous level. The current spread of Internet of Things (IoT) devices is causing an even greater acceleration, with the number of IoT devices online already outnumbering the number of people in the world. Within all of this data is information of potential value to many missions of the U.S. government. While IoT devices are plentiful, they tend to be inexpensive and are sometimes unreliable in their measurements. A prudent consumer of IoT data would be skeptical of any information gleaned from a single IoT sensor. This, however, is where the sheer volume of IoT devices in the world can provide a substantial benefit. The low cost of IoT sensors may yield relatively low performance, but also makes it feasible to collect large amounts of data over a sea of devices. While, for example, a single traffic sensor may frequently malfunction, fusing data across a large number of sensors improves robustness to unreliable measurements from any given device. A diverse array of sensor types, application areas, and data fusion methodologies form a large body of work in the academic literature. This report provides a survey of the recent literature on fusion techniques for IoT data, with an eye toward methods that may be interesting for U.S. government analysts, enabling them to augment their data most effectively and provide the highest possible force multiplier for their analysis products. The remainder of this report is organized as follows. Section 2 provides a brief background on the Internet of Things. Section 3 formally states the objective of the current study. The findings from several literature surveys are summarized in Section 4. Section 5 highlights recent results in relevant application areas. Finally, in order to get a set of findings that can be compared to each other, Section 6 outlines results in a particular application: indoor positioning. Section 7 provides a discussion of the implications of the current literature on future research and development. Section 8 summarizes and concludes the report.