Linking Sensed Images to an Ontology of Obstacles to Aid in Autonomous Driving

In this paper, we discuss the importance of recognizing and representing both stationary and moving obstacles for the purpose of autonomous driving as well as linking these representation to an ontology of obstacles to aid in deducing additional information about them. With the ability to access additional information about a sensed obstacle, an autonomous vehicle can better forecast where that obstacle can and can not be at a future time, and therefore be able to better plan its path to avoid collision with that obstacle. This paper describes work just recently begun at the National Institute of Standards and Technology in developing and incorporating an ontology of moving obstacles into the control of an autonomous vehicle to aid in path planning and obstacle avoidance.