Can We Recognize Atmosphere as an Agent?

Abstract

There are many environment control systems such as air conditioning systems for providing suitable environment to users. However, there are significant individual differences in the sense of suitable environments among users. Hence, a recommendation system indicating personally suitable environments for each user is extremely important. For this purpose, we are now developing an interactive human–machine interface system that dynamically recommends a suitable room environment for users. In this study, we investigated how the anthropomorphic appearance of the recommending system affected the satisfaction of the users. The results showed that the intellectual impression of the appearance was more important than the emotional impression when using the recommendation system.