An Approach for Utilizing Semantic Information in Autonomous Robot Control

摘要

This paper proposes a framework which utilizes semantic information of task world to realize robust task execution and monitoring by autonomous robots. Semantic information is indispensable for object recognition and handling in real world robotic tasks. We discuss the way of modeling of the task world and representation of robot plans. We are constructing an ontology and a prototype of proposed system for blocks-world task executed by a robot.