Vehicle vibration modelling via road profile reconstruction and simulation: coupled analysis of IRI, speed, and ride comfort for driving assistance

With the advancement of intelligent transportation and the “vehicle-road collaboration” concept, road roughness has become a key factor affecting ride safety and comfort. This study proposes a road elevation reconstruction method based on the International Roughness Index (IRI) and integrates it with the CarSim platform to simulate vehicle vibration responses under various speeds and IRI conditions. The weighted root mean square acceleration defined by ISO 2631 is adopted to evaluate ride comfort, and multiple regression, machine learning and deep learning models are employed to capture the nonlinear coupling between speed and roughness. Based on comfort thresholds, a driving strategy framework is developed to determine the optimal recommended speed. Field validation confirms that machine learning models effectively predict comfort levels, providing a reliable basis for intelligent driving assistance and pavement maintenance optimisation.