Huijun Deng

As an interactive medium between drivers and vehicles, in-vehicle virtual robots significantly impact user performance and experience. Among them, those robots that have anthropomorphic appearances are the mainstream. As an emerging human-vehicle interaction interface, Augmented Reality Head-Up Display (AR-HUD) still has not been used to showcase in-vehicle virtual robots. For this reason, it is currently unclear what preferences users have for the anthropomorphic appearance of virtual robots when using AR-HUD. In this work, we conducted two experimental online studies focusing on two aspects of anthropomorphic appearance: human similarity and morphological completeness. A total of 257 participants’ valid data sets were collected. Participants rated the in-vehicle virtual robot’s anthropomorphic appearance on five preference dimensions on Pleasure, Fear, Trust, Comprehensibility, and Acceptance dimensions. We found that users prefer appearances of virtual robots with medium human similarity and high morphological completeness, which aligns with the current theories on robot appearance. Our future research will focus on developing and validating design principles for the appearances of in-vehicle virtual robots.

While automated driving has attracted more and more attention, human drivers still play an important role. Drivers must be aware of their surroundings and potential danger to take over the vehicle’s control immediately when the automation system cannot function, which could otherwise pose a road safety risk. The transparent design of the human–machine interface can accurately convey the perceived information of the machine to help drivers understand the dynamic environment. We designed the scenario where the rear vehicle is about to overtake the ego vehicle and has a dangerous tendency to collide with the ego vehicle by stepping on the lane line. We analyzed the information shared by the intelligent system in this scenario to make “transparent” the intelligent system. According to the requirements of the task model of transparency and the three stages of situation awareness, we formed three designs. Then we analyzed the availability, situation awareness, and workload of different designs. Results showed that the transparency design according to the shared situation awareness stage can have a positive impact on the usability of the intelligent system. Therefore, the effective transparency-based design in automated driving can enhance the user’s understanding of the system and improve their experience.