A New Epistemic Fluent within the Situation Calculus Using Linguistic Variables

An alternative epistemic fluent is proposed to solve the important issues of reasoning and sensing in intelligent control for autonomous agents. The approach treats interval-valued fluents within the situation calculus John Funge proposed as linguistic variables, the members of which are fuzzy sets defined on the interval. The new approach can integrate discrete-valued epistemic fluents with continuous-valued epistemic fluents and make cognitive model of autonomous agent much simple. The paper describes how to represent non-deterministic actions in the situation calculus based on epistemic fluent K and how to solve the potentially uncountable number of possible worlds in agent sensing by using this new approach. Finally the new approach was applied to create autonomous characters which can counteract with player characters in game Quake 2.