Indexical Knowledge in Robot Plans

Robots act upon and perceive the world from a particular perspective. It is important to recognize this relativity to perspective if one is not to be overly demanding in specifying what they need to know in order to be able to achieve goals through action. In this paper, we show how a formal theory of knowledge and action proposed in (Lesperance 1989) can be used to formalize several kinds of situations drawn from a robotics domain, where indexical knowledge is involved. Several examples treated deal with the fact that ability to act upon an object does not require de re knowledge of the object or its absolute position; knowledge of its relative position is sufficient. It is shown how the fact that perception yields indexical knowledge can be captured. We also point out the value of being able to relate indexical knowledge and objective knowledge within the same formalism through an example involving the use of a map for navigation. Finally, we discuss a problem raised by some higher-level parametrized actions and propose a solution.