Synthetic Characters with Emotional States

A new trend that has emerged in the field of applied artificial intelligence in the past few years is the incorporation of emotion emulation mechanisms in situated agent systems. Several researchers in the field of psychology and neural science agree that emotion is an essential aspect of human intelligence, diverging from the traditional treatment of emotion as something that inhibits rational thinking. This argument has also influenced the area of artificial intelligence and especially the young, yet vibrant field of Intelligent Virtual Agents, where it has become accepted that emotion is a key issue to achieve believable agent behaviour. The increased interest for emotions has resulted in several computational emotional models having been presented, with diverse approaches, generically classified into cogn itive and non-cognitive, inspired by areas such as neural science. Hybrid approaches have also appeared, combining both cognitive and noncognitive elements in an attempt to accurately describe the inner workings of emotional mechanisms. Such a hybrid model is the one we adopt in this paper, aiming to apply it to SimHuman, an intelligent virtual agent platform, in order to introduce emotion awareness and affective behaviour to the agent architecture.