An Architecture for Cognitive Modeling to Support Real-Time Adaptation and Motivational Responses in Video Games

摘要

Currently, there are tremendous advances in gaming technologies to improve physical realism of environments and characters. However, game characters are still lacking in the cognitive realism, thus video games and game development tools need to provide functionality to support real-time adaptation and appropriate response to internal motivations. In this paper we propose an architecture for cognitive modeling based on well-known cognitive architectures like ACT-R and Soar. It describes a methodology for developing a behavioral system and it emphasizes an ethological sensing modeling to simulate more realistic behaviors. This methodology serves to develop a component-based system that could improve the ability to create intelligent virtual agents. In this phase of implementation, we present preliminary results of modeling behavior of a fish character in an undersea world using a game engine.