A logic for specifying agent actions and observations with probability

Abstract

We propose a non-standard modal logic for specifying agent domains where the agent’s actuators and sensors are noisy, causing uncertainty in action and perception. The logic is multi-modal, indexed with actions; the logic is also augmented with observation objects to facilitate knowledge engineers dealing with explicit observations in the environment, and it includes a notion of probability. A tableau method is provided for proving decidability of the proposed logic. It is our conjecture that the tableau rules are complete with respect to the semantics. The proof does not yet exist, however, we discuss the current approach of the proof and provide some examples to motivate our conjecture.