Anomaly Detection in Simulated Vehicle Dynamics Using BeamNG.tech and the TEDA Framework

Simulations are crucial in the automotive industry for enhancing safety and cost efficiency, especially given the growing complexity of vehicle systems. This study proposes a methodology that integrates the BeamNG.tech simulation platform with a data collection system and the Typicality and Eccentricity Data Analytics (TEDA) framework to detect anomalies in vehicle speed data, facilitating driver behavior analysis. A case study was conducted where a driver navigated a predefined route while performing maneuvers with intentional speed variations. The TEDA framework successfully identified 1,110 outliers out of 6,388 speed samples, representing approximately 17.38% of the total data. The proposed methodology enables the detection of atypical driving behaviors, demonstrating its potential to contribute to the development of advanced driver assistance systems and improve safety technologies.