Road Anomalies Mapping Using Sensor Data from Mobile Devices

Road anomaly detection is essential for enhancing vehicle safety and comfort by identifying irregularities such as bumps and potholes. This study utilizes accelerometer and GPS data from a smartphone to detect road anomalies. It employs threshold-based methods to analyze the data, initially separating the axes and calculating the standard deviation and acceleration magnitude. To improve accuracy, a dynamic rolling window approach is introduced, which automatically detects bumps by adjusting the window size based on local speed. Additionally, an SVM (Support Vector Machine) algorithm was analyzed and compared to the threshold method, providing a robust comparison of anomaly detection techniques and validating the effectiveness of the threshold-based approach.