Estimation of Motion Sickness in Automated Vehicles using Stereoscopic Visual Simulation

Automation of driving leads to decrease in driver agency, and there are concerns about motion sickness in automated vehicles. The automated driving agencies are closely related to virtual reality technology, which has been confirmed in relation to simulator sickness. Such motion sickness has a similar mechanism as sensory conflict. In this study, we investigated the use of deep learning for predicting motion. We conducted experiments using an actual vehicle and a stereoscopic image simulation. For each experiment, we predicted the occurrences of motion sickness by comparing the data from the stereoscopic simulation to an experiment with actual vehicles. Based on the results of the motion sickness prediction, we were able to extend the data on a stereoscopic simulation in improving the accuracy of predicting motion sickness in an actual vehicle. Through the performance of stereoscopic visual simulation, it is considered possible to utilize the data in deep learning.