Bluetooth-Based Vehicle Counting: Bridging the Gap to Ground-Truth With Machine Learning

Abstract

Traffic flow, number of vehicles passing a particular point over a given period of time, is an essential indicator for evaluating the performance and condition of road networks, detecting congestion, and predicting traffic trends. Accurate and reliable measurement of traffic flow in urban roads is challenging due to the dynamic nature of intersection signals and comes with high equipment and maintenance cost. WaveTraf is a Bluetooth-based Intelligent Traffic System solution widely deployed in the State of Qatar which detects and monitors the movement of Bluetooth-enabled devices anonymously using their unique MAC addresses. Systems such as WaveTraf allow for real-time, low-cost, scalable and non-intrusive traffic flow measurement; however, they could suffer from low detection and sampling rates leading to uncertain and unreliable estimates. In this research, we investigate various machine learning techniques such as Random Forrest, Support Vector Regression Machines and XGBoost to model the relationship between the ground-truth traffic flow based on video cameras and Bluetooth-based traffic flow. We utilized these techniques to enhance the dependability of Bluetooth-based traffic flow measurements, making it a more desirable and cost-effective solution for real-time traffic flow measurement.