Estudio, modelado y desarrollo de inteligencias virtuales en entornos simulados persistentes

Abstract

The confluence of 3D virtual worlds with social networks imposes to software agents, in addition to his conversational functions, the same behaviors as those common to human-driven avatars. In this thesis we explore the possibilities of the use of metabots (metaverse robots) in virtual 3D worlds and we introduce a new cognitive model based on cooperation of metabot’s capabilities on a working contextual memory over which environment semantic knowledge is built. Connecting through an avatar to a virtual world which is semantically isomorphic to reality allows a convergence to take place between the concepts used by humans and those used by machines. By defining a common ontology we can enhance the meaning of the communicative act. In the virtual world, both autonomous avatar and human driven avatar act as an active entity that has information about itself and its surrounding environment. As actor in a 3D space, its activity will be similar to that performed by humans in the physical world and, therefore, it needs semantics similar to those used by humans to describe our activities. This thesis describes the fundamental capabilities that any Virtual Embodied Agent on persistent spatial-temporal environments must fulfill, develops an architecture to withstand those capabilities requirements and proposes a cognitive model to allow the metabot to interact with any element present on the virtual environment. Finally the results of experimentations will be shown performed to test the theoretical model as well as the implemented architecture defined along this thesis.