FoGBAT: Combining Bluetooth and GPS Data for Better Traffic Analytics

Congestion is a major problem in many cities. In order to monitor and manage traffic, a number of different sensor types are used to collect traffic data. This includes GPS devices in the vehicles themselves as well as fixed Bluetooth sensors along the roads. Each sensor type has advantages and disadvantages. Where GPS has a wide coverage of the road network, Bluetooth sensors gather data from a much higher number of vehicles. In this paper we present Fog BAT, a system that combines GPS data with Bluetooth data. The goal of the system is to retain the advantages of both. We show how the data types are aligned to ensure that data from each sensor type is related to the exact same part of the road network and cover the same time period. Using very large real-world data sets, we use the system to compare travel speeds based on each data type, and how the use of both data types simultaneously can improve the accuracy of computed travel speeds and congestion levels.