Assessing Passengers' Motion Sickness Levels Based on Cerebral Blood Oxygen Signals and Simulation of Actual Ride Sensation

Abstract

(1) Background: After motion sickness occurs in the ride process, this can easily cause passengers to have a poor mental state, cold sweats, nausea, and even vomiting symptoms. This study proposes to establish an association model between motion sickness level (MSL) and cerebral blood oxygen signals during a ride. (2) Methods: A riding simulation platform and the functional near-infrared spectroscopy (fNIRS) technology are utilized to monitor the cerebral blood oxygen signals of subjects in a riding simulation experiment. The subjects' scores on the Fast Motion sickness Scale (FMS) are determined every minute during the experiment as the dependent variable to manifest the change in MSL. The Bayesian ridge regression (BRR) algorithm is applied to construct an assessment model of MSL during riding. The score of the Graybiel scale is adopted to preliminarily verify the effectiveness of the MSL evaluation model. Finally, a real vehicle test is developed, and two driving modes are selected in random road conditions to carry out a control test. (3) Results: The predicted MSL in the comfortable mode is significantly less than the MSL value in the normal mode, which is in line with expectations. (4) Conclusions: Changes in cerebral blood oxygen signals have a huge correlation with MSL. The MSL evaluation model proposed in this study has a guiding significance for the early warning and prevention of motion sickness.

Keywords: Bayesian ridge regression (BRR); Graybiel scale; cerebral oxygenation signals; functional near-infrared spectroscopy (fNIRS); motion sickness; vehicle test.

Figure 1

Experimental flow design based on riding simulator.

Figure 2

Experimental equipment and testing process. (a) ride simulation experiment; (b) simulation platform; (c) fNIRS device.

Figure 3

Normality test of signals (one subject).

Figure 4

Pearson correlation analysis and significance test for independent and dependent variables.

Figure 5

Four-in-one plot for residual analysis. (a) Residual–observation plot; (b) Residual histogram; (c) Residual–fitted value scatter plot; (d) Probability–residual plot.

Figure 6

Cross-validation of predicted motion sickness value and Graybiel score.

Figure 7

Vehicle test and random route area. (a) subject; (b) experimenter; (c) running route.