Virtual Agents’ Self-Perception in Story Telling

The goal of this work is self-perception modeling of autonomous agents in virtual story telling. It’s inspired from work of psychologists and neuro-physiologists. From psychology, we use fuzzy cognitive maps (FCM) to model and implement believable agents’ behaviours. These cognitive maps allow us to give not only sensation but also perception, in the sense that our agents perceive environment in function of their inner states or emotions. From neuro-physiology, we implement the idea that movement is simulated in the cortex before it is performed in real world. Virtual agent’s self-perception is the ability to simulate different behaviours in its own imaginary space before acting in “real” world. This self-perception implemented by “simulation in the simulation” is one of the keys for the autonomy of virtual entities’ decision.