Validity Quantification of Driver-in-the-Loop Simulation in Motorsport

Abstract

Driving simulators are indispensable tools to be competitive in motorsport, for drivers as well as engineers. Fidelity and validity of a driver-in-the-loop simulator determine its utility for car setup development, drivers' training and race strategy investigations. The conclusions drawn from race preparations at a driving simulator take its validity at the vehicle's dynamic limits as a basis. A high level of simulator fidelity does not necessarily imply validity of research outcomes. Actuators, ergonomics and screen size as well as track model, vehicle model and motion cueing algorithms could influence simulator validity. Whereas the impact of track and vehicle model can be quantified, the impact of simulator motion on simulator validity is not yet holistically defined as objective data. Therefore, a method which quantifies the overall validity and the impact of individual simulator components is of high interest for further development. The methodology to quantify simulator validity is based on driving style identification. A method was introduced earlier in our department to categorize race drivers, driving at the limit of a vehicle's dynamic capabilities. From a motorsport engineer's point of view the overarching objective of simulator development is to have minimum deviation in driving style between track and simulator tests. Race drivers' driving style is defined, but not readily apparent, by their interactions with steering wheel and pedals. Recorded data of simulator and track operation is processed to calculate metrics during specific vehicle states. In this work the resulting driver metrics are further processed to driving style deviation metrics which describe discrepancies between race track and simulator operation. An evaluation of the derived metrics allows simulator validity quantification. The impact of motion stimuli on simulator validity is compiled using the introduced method to prove its relevance. As a result, the here presented method serves as a measure of motorsport simulator validity. Additionally, the method allows to quantify driving style deviation at variable simulator setups. The impact of various simulator components on simulator validity can be analyzed consequently. A limitation of the developed methodology is that the driver metrics are only validated for the classification of professional race drivers, driving the cars at the limit of their dynamic capabilities. Furthermore, validated track and vehicle models are mandatory requirements to evaluate the impact of motion stimuli on absolute validity of the simulator. Knowledge about the impact of various components on simulator validity will provide objective guidance for future driving simulator development. In this particular case, research on evaluation and optimization of motion cueing algorithms will be carried out which is motivated by the obtained findings. Special focus will be on the motion stimuli while driving the simulated vehicle close to its dynamic limits.