Efficient, realistic NPC control systems using behavior-based techniques

Abstract

Computer games are an application in which the perception of intentionality is often more important than intentionality itself. Players often attribute more intelligence to non-player characters than is actually warranted. Therefore, rather than investing more AI complexity in the system, we propose an approach which refines and extends the finite-state machine techniques already prevalent in commercially available games. Taking advantage of the similarities between robots and NPCs in dynamic environments, we describe two NPC control systems that uses behavior-based techniques.