Towards cosmopolitan robots: intelligent navigation in extended man-made environments

Abstract

In the past, mobile robots have been constrained to operate in either an indoor or
\nan outdoor environment, not both. Special purpose representations and ad hoc sensor
\ntechniques geared towards tasks of narrow focus have dominated these efforts. It is the
\npurpose of this dissertation to lead towards the development of a more cosmopolitan
\nrobot; one whose domain of interaction is not as restricted as these previous attempts.
\nThe Autonomous Robot Architecture (AuRA) has been developed to meet these
\nchallenges. A "meadow" map, used for global path planning and containing embedded a
\npriori knowledge to guide sensor expectations, serves as the robot's long term memory.
\nA layered short term memory based on instantiated meadows represents the currently
\nperceived world. A hierarchical path planner produces a global path free of collisions
\nwith all modeled obstacles. Schema theory is extended to include the mobile robot domain and serves as the
\nprincipal theoretical framework. The schema-based path execution system handles unexpected
\nand dynamic obstacles not present in the robot's world model. This motor
\nschema-based navigation system produces reactive/reflexive behavior in direct response
\nto sensor events. In addition, new techniques in the treatment of robot uncertainty which
\nexpedite sensory processing are presented. These include the use of a spatial error map
\nwith associated error growth and reduction techniques. Several computer vision sensor strategies have been developed for use within AuRA.
\nThese include a fast line-finding algorithm, a fast region segmentation algorithm, and a
\ndepth-from-motion algorithm. Experiments using our mobile vehicle HARV demonstrate
\nthe use of these vision algorithms for navigational purposes. Schema-based navigation
\nusing ultrasonic sensing is also demonstrated experimentally.