Combining Tracking Data of Multiple Sources for Increased Precision

The challenges in tracking passive mobile devices involve the lack of precision in estimating device positioning and the difficulty in providing a unique identification of the devices. In this sense, the combination of different approaches such as Wi-Fi 802.11 tracking, Bluetooth, among other sources (LiDAR, Cameras), can provide greater precision in estimating accuracy and uniquely determining devices. These data are typically collected at different nodes (e.g., access points, Bluetooth beacons) and centralized in a sink that must correlate the different sources to estimate the devices' positioning at a point in time.

This thesis aims to facilitate the process of calculating the precision of a device location by delivering a service that allows entering spatial data and sensor-types of multiple wireless trackers, which will then return the precision percentage of those devices. The correlation and analysis-engine should also predict the possible outcome of an event based on past events. Henceforth, the data attributes needed to calculate will need to be defined beforehand to achieve this goal. Initially, the spatial coordinates (X, Y), a timestamp, and a sensor-type will be taken as an input for the engine. They should return a precision for the desired device position, allowing further calculations (e.g., the distance between people in a shop) to be more accurate considering the respective precision value of the devices’ position. Considering the different available sensor-types, each sensor-type (e.g., Bluetooth and Wi-Fi) may have the X- and Y-axis.