Findings from subjective evaluations and driver ratings of vehicle dynamics: steering and handling

This paper investigates subjective assessments (SA) of vehicle handling and steering feel tests, both numerical and verbal, to understand drivers’ use of judgement scales, rating tendencies and spread. Two different test methods are compared: a short multi-vehicle first-impression test with predefined-driving vs the standard extensive single-vehicle free-driving tests, both offering very similar results but with the former saving substantial testing time. Rating repeatability is evaluated by means of a blind test. Key SA questions are identified by numerical subjective assessment autocorrelations and by generating word clouds from the most used terms in verbal assessments, with both methods leading to similar key parameters. The results exposed in this paper enable better understanding of SA, allowing improving the overall subjective testing and evaluation process, and improving the data collection and analysis process needed before identifying correlations between SA and objective metrics.     

A method to find correlations between steering feel and vehicle handling properties using a moving base driving simulator

There have been several investigations to find out how drivers experience a change in vehicle-handling behaviour. However, the hypothesis that there is a correlation between what the driver perceives and vehicle- handling properties remains to be verified. To define what people feel, the human feeling of steering systems was divided into dimensions of perception. Then 28 test drivers rated different steering system characteristics of a semi-trailer tractor combination in a moving base-driving simulator. Characteristics of the steering system differed in friction, damping, inertia and stiffness. The same steering system characteristics were also tested in accordance with international standards of vehicle-handling tests resulting in characteristic quantities. The instrumental measurements and the non-instrumental ratings were analysed with respect to correlation between each other with the help of regression analysis and neural networks. Results show that there are correlations between measurements and ratings. Moreover, it is shown that which one of the handling variables influence the different dimensions of the steering feel.     

Links between subjective and objective evaluations regarding the steering character of automobiles

This work presents an experiment on the relationships between subjective and objective evaluations of vehicle handling. Ten cars were examined objectively in several open-loop driving dynamics manoeuvres and subjectively by test persons in typical traffic situations. Results are extracted from a stationary test (the Slowly Increasing Steer Test), and a dynamical test (the Frequency Response Test). The subjective measurements are obtained from drivers on a rural road course via a questionnaire, which was developed to separately investigate the quantity level perception, the so-called “Niveau”, and the more qualitative “Liking”. These subjective “measurements” are embedded into a two-channel definition of “Steering Comfort” as a genus for “Steering Discomfort” and “Character”. The article concentrates on developing a statistical method for the consideration of correlations amongst the subjective/objective data. In doing so, the variance in example subjective Niveau ratings can be significantly explained by several objective quantities. Indicators for co-domains of validated discomfort characteristics and hints for endeavouring character Liking ranges are detected.     

A polynomial chaos approach to the analysis of vehicle dynamics under uncertainty

The ability of ground vehicles to quickly and accurately analyse their dynamic response to a given input is critical to their safety and efficient autonomous operation. In field conditions, significant uncertainty is associated with terrain and/or vehicle parameter estimates, and this uncertainty must be considered in the analysis of vehicle motion dynamics. Here, polynomial chaos approaches that explicitly consider parametric uncertainty during modelling of vehicle dynamics are presented. They are shown to be computationally more efficient than the standard Monte Carlo scheme, and experimental results compared with the simulation results performed on ANVEL (a vehicle simulator) indicate that the method can be utilised for efficient and accurate prediction of vehicle motion in realistic scenarios.     

Virtual test rig to improve the design and optimisation process of the vehicle steering and suspension systems

A virtual test rig is presented using a three-dimensional model of the elasto-kinematic behaviour of a vehicle. A general approach is put forward to determine the three-dimensional position of the body and the main parameters which influence the handling of the vehicle. For the design process, the variable input data are the longitudinal and lateral acceleration and the curve radius, which are defined by the user as a design goal. For the optimisation process, once the vehicle has been built, the variable input data are the travel of the four struts and the steering wheel angle, which is obtained through monitoring the vehicle. The virtual test rig has been applied to a standard vehicle and the validity of the results has been proven.