Combining Uncertainty and Description Logic Rule-Based Reasoning in Situation-Aware Robots

Abstract

The paper addresses how a robot can maintain a state repre-sentation of all that it knows about the environment over time and space, given its observations and its domain knowledge. The advantage in combining domain knowledge and obser-vations is that the robot can in this way project from the past into the future, and reason from observations to more general statements to help guide how it plans to act and interact. The difficulty lies in the fact that observations are typically uncer-tain and logical inference for completion against a knowledge base is computationally hard.