People tend to feel uneasy when the driver is not there with self-driving cars, even if they know it's safe. Therefore, it is important to know the passengers' level of comfort from their facial expressions. In this study, a model of an autonomous vehicle runs on a virtual road set up in the laboratory. When the autonomous vehicle detects an obstacle, the car avoids disaster with Energy optimal control (EOC). To be considered with digital twin, the same check is also carried out on the traffic flow simulator. We develop a method to measure the sense of security from facial expressions of vehicle occupants (subjects) before and after the car avoids danger using machine learning and we study the EOC control function that improves the sense of security.
ABE et al. (Wed,) studied this question.