An Intelligent World Model for Autonomous Off−Road Driving

This paper describes a world model designed to act as a bridge between multiple sensory inputs and a behavior generation (path planning) subsystem for off−road autonomous driving. It describes how the world model map is built and how the objects and features of the world are represented. The functions used to maintain the model are explained and the sensors and sensory processing used to provide data for this application are discussed. The paper includes examples of integrating and fusing sensory data from multiple sources into the world model map. The representation is being developed for the Army’s Demo III autonomous driving experiment, which is an on−going research project. The paper concludes with a discussion of future research directions.