Synthetic emotension building believability

Abstract

We present our first steps towards a framework aiming at increasing the believability of synthetic characters through attentional and emotional control. The framework is based on the hypothesis that the agent mind works as a multi-layered natural evolution system, regulated by bio-digital mechanisms, such as synthetic emotions and synthetic attention, that qualitatively regulate the mind endless evolution. Built on this assumption, we are developing a semi-autonomous module extending the sensor-effector agent architecture, handling the primary emotensional aspects of the agent behavior thus, providing with the necessary elements to enrich its believability.