Improved Predictive Model of Drivers’ Subjective Perception of Vehicle Reaction under Aerodynamic Excitations

Artikel i vetenskaplig tidskrift, 2023

In vehicle development, rating vehicle reactions to external disturbances such as aerodynamic excitations are important for improving safety and comfort of passengers. Vehicle motion reactions under such conditions are dependent on both disturbance and drivers’ steering actions. A predictive model has been developed to correctly anticipate the drivers’ ability to identify unexpected external disturbances for straight-line, high-speed driving in a driving simulator. The measured variables were band-pass filtered to desired frequency ranges. Excess yaw and roll velocities, defined as the difference between actual rotations and rotations predicted with a dynamic model from the cause of actual steering, were calculated. The standard deviations of the measured variables in a time window around disturbances were used in a regression model to predict the driver responses. Replacing actual rotations with excess rotations reduced the importance of steering input as a predictor by approximately 2/3, thus improving the accuracy of the predictive model. The model showed the change in driver sensitivity to rotations at different frequency intervals. It also showed that only driver input in around 1 ∼ 2 Hz reduces driver sensitivity and that drivers are not necessarily sensitive to high rotational accelerations, but rather to large differences between actual vehicle yaw and roll and expected vehicle yaw and roll responses from the steering input The result from this study were later compared to succeeding on-road tests which confirmed that the predictive model was improved with the use of excess motion variables.