The development of novel computer interfaces has led to the possibility of integrating inputs from multiple individuals into a single avatar, fostering collaboration by combining skills and sharing the cognitive load. However, the collaboration dynamic and its effectiveness may vary depending on the individuals involved. Particularly in scenarios where two individuals remotely control a robotic avatar without the possibility of direct communication, understanding each other’s characteristics can result in enhanced performance. To achieve this, it is essential to ascertain if individuals can discern their partner’s characteristics within the merged embodiment. This paper investigates the accuracy with which participants can distinguish between two different collaborating partners (one attempting to lead and one attempting to follow) when sharing control of a robot arm during a block pick-and-place task. The results suggested that participants who changed their roles according to the different roles of the two partners achieved the highest discrimination rates. Furthermore, participants changed their movements through the trials, adapting their actions to their preferred approach. This research provides insights into the factors determining individuals’ ability to understand partner characteristics during control of collaborative avatars.

Humans have one own body. However, we can share a body with two different persons in a virtual environment. We have developed a shared body as an avatar that is controlled by two persons' actions. Movements of two subjects were continuously captured, and integrated into the avatar's motion with the ratios of 0:100,25:75, 50:50, 75:25, 100:0. They were not aware of integration ratios, and asked to reach cubes with the right hand. They felt body ownership and sense of agency to the shared avatar more when the responsible ratio was higher. The reaching path of the avatar's hand was shorter in the shared body condition (75:25) than the single body condition (100:0). These results suggest that we have some of body ownership and sense of agency to the shared body, and the task performance is improved by the shared body.

Pathogenic infection is usually invisible. Thus, we hardly care of preventing it in the daily life. We aimed to develop a virtual reality system to experience bidirectional infections using vision and tactile sensations; Users are infected by an avatar, and they do infect the avatar. When the avatar coughs toward a user, a cloud of visualized pathogen-droplets jet out and strike the user and objects on a table. At the same time, vibrations are presented to the user's chest. When the user coughs toward the avatar, the cough was detected by a throat microphone, and a cloud of visualized pathogen jets out. The avatar tries to avoid it unpleasantly, but infected. He is getting sick in bad face color. This system enables us to experience pathogenic infection with multimodal sensations in both perspectives (infected and infecting). Thus, it may contribute to increasing user's knowledge of infection and to facilitating prevention behaviors such as wearing a hygiene mask.