Tsoutsanis A., Yannis G., “Classifying Vehicle Cornering Behavior using Mobile Sensor Data”, Proceedings of the 13th Symposium of the European Association for Research in Transportation (hEART), Munich, 10-12 June 2025.

This paper presents a methodology to classify cornering behavior in vehicles using data from mobile devices equipped with accelerometers and gyroscopes. Data was captured from a fixed-position device on a vehicle, collecting accelerometer (3 axes), gyroscope (3 axes), and GPS speed. Two driving modes were analyzed: Normal and Aggressive, with data collected from over 100 trips. The paper describes aligning the device’s coordinate system with the vehicle’s, preprocessing data through peak detection, augmentation, and downsampling, and applying a four-layer Long Short-Term Memory (LSTM) network to classify cornering as normal or harsh. Experimental results demonstrate the model’s effectiveness, achieving 84% accuracy in distinguishing driving cornering behaviors.