Abstract
Keywords:vehicle dynamics, steering comfort, model-based testing, open-loop tests, subjective-objective correlations
Hereafter, an integrated method to investigate the correlation between objective and subjective evaluations of steering torque aspects is described in order to front-load and optimize development efforts. It deals with the links between subjective impressions the driver experiences while guiding a vehicle and objective measures from open-loop vehicle dynamics tests. This is accomplished via an experiment with subjects evaluating different steering torque settings in one test vehicle on a rural road profile. Therefore, the car is equipped with a steering torque actuator, which is in turn controlled by a combined Extended Single Track-Steering System Model.
Hereby, the vehicle model provides the inputs for the steering system which in turn computes the steering torque. With respect to the on-centre regime, special attention is paid to the friction modelling. By means of the combined model, a "fingerprint" of the vehicle characteristics can be identified. The method involves vehicle test-rig and open-loop driving manoeuvres as inputs for the model identification. Measurement data collected from those real vehicle tests is fed into an optimization of model parameters. Once the fingerprint is optimized, component parameters can easily be changed virtually, their effects on the test characteristics studied in offline simulations and assessed subjectively in reality.
Herein, these tools are implemented to validate objective metrics from two open-loop driving dynamics manoeuvres, Weave and Slowly Increasing Steer Test. The experimental settings are made up of real cars and their derivates with altered component properties. Test drivers subjectively rate those settings in on- and off-centre questions. One goal is to disclose whether the evaluation criteria are steering discomfort or character topics under the genius of steering comfort. Therefore, the implemented models span a wide range in the objective quantities. On the subjective side, a questionnaire is presented that distinguishes between a "Niveau" and "Liking" scale in order to investigate the relations between the perception level and taste.
Almost every subjective criterion emerges to be a question of steering character. On the objective side, steering torque gradient, friction deadband from Weave and torque magnitudes and gradients from Slowly Increasing Steer Test are found to be key objective characteristics for the description of steering comfort. Hence, the method bridges the gap between virtual and real vehicle steering comfort design by gearing components, overall vehicle behaviour and subjective-objective evaluations.