Road Anomalies Detection Using Low-Cost Sensors and Machine Learning

In this paper we introduce SVRUM: a cost-effective sensor platform to be mounted on Vulnerable Road User (VRU) vehicles (bicycles, e-bikes, kicks scooters) for the detection of road anomalies. SVRUM consists of a 3-axis accelerometer, a short-range sonar sensor, and a GPS module, which are connected to an Arduino board. We used SVRUM to collect data for eight different types of road anomalies (now publicly available to the community) and test various data analysis techniques, including machine learning algorithms, to identify road anomalies. The results are auspicious and demonstrate the potential of SVRUM in enhancing road safety for VRUs.