Predictive model of perceived driving stability at high speeds under aerodynamic excitations

Abstract: The automotive industry is continuously advancing towards more energy-efficient vehicle designs.  Streamlined vehicles have low aerodynamic drag but have the potential to be unstable when  exposed to external excitations such as unsteady aerodynamic forces created by the flow of air  around them.  Before signing off a new vehicle for production, several on-road test scenarios are  conducted by professional drivers to evaluate the performance. Finding vehicle instabilities and  proposing solutions to problems during late phases of development is challenging and costly. The objective of this thesis is to correlate and predict the driver's subjective evaluation of high  speed straight-line driving stability with measurable quantities in early design phases.  In this work, substandard straight-line drivability was investigated on-road using different  aerodynamic devices for generating high rear lift and asymmetric aerodynamic forces. These  aerodynamic devices were then paired with stabilizers, called side-kicks, which helped to define  the flow separation and improved the drivability of the tested vehicle. Vector plots of the  mean and standard deviation of lateral acceleration, yaw velocity, steering angle, and steering  torque were used to understand vehicle behavior for the paired configurations and relate  to the difference of subjective evaluation of drivability within each pair. The ride diagram  was used to separate the presence of transient behavior and study its impact on subjective  evaluation. The qualitative assessment of the resulting trends agrees well with the subjective  evaluation of the driver. Following this, experimental trials were conducted in driving simulators and on-road, in order to  have an in-depth understanding of drivers' subjective evaluation and responses to external  excitations. Both common and professional test drivers were involved in the study. The results  provided insight into the excitation frequencies and amplitudes of interest. From the test  data, mathematical models were generated that can predict the drivers' subjective evaluation  after experiencing induced external excitations. The outcome showed the impact of drivers'  steering on their subjective evaluations towards these excitations. The on-road study revealed  that higher roll and longitudinal noises reduce the drivers' sensitivity to external excitations.  Headwind magnitude and lateral motion in a certain frequency range experienced by the  human upper body contribute to drivers' identification of excitations. The resulting predictive  model can be used to pinpoint the time of occurrence of observable aerodynamic excitations  and provides their characteristics in early development phases. Since the models represent  measurements from the cabin, they should be valid for different vehicles.