Experimental evaluation of an optimization-based motion cueing algorithm. (April 2019)

• Record Type: Journal Article
• Authors: Ellensohn, Felix; Venrooij, Joost; Schwienbacher, Markus; Rixen, Daniel
• Is Part Of: Transportation research. Volume 62(2019)
• Page Start: 115
• Page End: 125
• Publication Date: 2019-04
• Subjects: Motion cueing algorithms -- Optimization-based -- Continuous rating; Automobile drivers -- Psychology -- Periodicals; Automobile driving -- Psychological aspects -- Periodicals; Transportation -- Psychological aspects -- Periodicals; 629.283019
• DOI: 10.1016/j.trf.2018.12.004
• Languages: English
• ISSNs: 1369-8478

Abstract

Highlights: Evaluation of an optimization-based motion cueing algorithm (MCA). Implementation of the continuous rating method in a driving simulation experiment. Potential of the novel MCA compared to a state-of-the-art MCA for a 9-DoF simulator. Development of a rating model to predict perceived motion incongruencies. Abstract: This paper describes a global optimization scheme that is employed as a motion cueing algorithm (MCA) for a 9 degrees of freedom (DoF) driving simulator. The paper describes the evaluation of the MCA in an experiment with 35 participants. Herein, the MCA is compared to a commercial, state-of-the-art, optimization-based MCA with the goal to illustrate the potential improvements in motion cueing quality that the novel MCA could provide. The experiment design includes a continuous rating method, where the participants continuously evaluate deviations between expected and perceived motions. Rating results confirm the potential of this novel MCA approach. Furthermore, the ratings are used to train a linear rating model. Here, the objective deviations between the physical motions of the simulator and the simulated vehicle functions as input. The model approximates the human perceptual system and the human rating procedure to produce a modelled rating. The proposed model achieves high correlations to the reference rating with the training set. However, it is questionable whether the linear rating model can be applied in general as the model does not accurately capture the ratings in a testing-set.