DriveBlue: Traffic Incident Prediction through Single Site Bluetooth

Transportation authorities, and traffic management aim to have a free-flow traffic on highways as well as in cities. This paper proposes DriveBlue, a real time system that uses Bluetooth adapters in a single site to predict traffic incidents in various conditions (e.g. congestion), and inform the responsible authorities. DriveBlue leverages the capabilities of co-located Bluetooth adapters to monitors vehicles on roads, categorize them, analyze their behavior over time, and reports any suspicious change. Moreover, DriveBlue has a simple design that is scalable using the infrastructure currently deployed by transportation authorities. In our experiments, we collected real data from highways to evaluate the feasibility of using single site Bluetooth adapters for transportation applications. DriveBlue spotted almost 10 devices per minute, and showed an accuracy of approximately 80% in detecting vehicle´s motion.