Extracting Driving Styles from Automotive Sensor Data to Develop Personas

Creating accurate and insightful personas for automotive applications requires understanding the diverse driving styles that emerge from sensor data. This paper introduces initial results of our methodology for developing personas based on the analysis of automotive sensor data, which captures various driving behaviors. We present a comprehensive process that begins with the extraction of driving events from sensor data, followed by the identification of distinct driving styles through K-Means clustering. Our approach is innovative in its use of dynamic, real-world driving data as opposed to static or direct user interaction data commonly employed in persona development. This allows for a deeper understanding of driver behaviors, which are indirectly inferred from sensor data, thus providing a foundation for creating detailed personas.