Route Decision System Associating with Facial Expression

Abstract

In this paper the case in which the way to an instructed destination is given by linguistic terms is considered as an example of human and computer interaction. Two models are considered. One is a route decision model that understands instructions are goes to the destination. The other is a facial expression model that expresses emotions according to the situation that the route decision model falls into. These two models are implemented on a computer. In the route decision model fuzzy sets are employed in order to deal with fuzziness inherent in linguistic terms. A degree of matching is definied to compare a fuzzy set representing the meaning of linguistic terms and a fuzzy set representing information on visual scene. The route decision model proceeds on its way by the use of the degree of matching. In the facial expression model emotions are inferred from the situation of the route decision model by the use of the fuzzy reasoning. Facial expressions are shown based on values of 6 variables related to eyes, eyedrows and mouth obrained by the use of a neural network model. Some simulation results show that the route decision model can reach an instructed destination in spite of fuzzy and uncertain instructions, the model sometimes loses its way as human does, and that by the facial expression model it becomes clear what emotions the route decision model has when it reaches the destination or loses its way.