Autonomous discovery and learning by a mobile robot in unstructured environments

Conference · Thu Dec 31 23:00:00 EST 1987

OSTI ID:6265108

Pin, F G; de Saussure, G; Spelt, P F; Barnett, D L; Killough, S M; Weisbin, C R

This paper presents recent research activities at the Center for Engineering Systems Advanced Research (CESAR) in the area of autonomous discovery and learning of emergency and maintenance tasks in unstructured environments by a mobile robot. The methodologies for learning basic operating principles of control devices, and for using the acquired knowledge to solve new problems with conditions not encountered before are presented. The algorithms necessary for the robot to discover problem-solving sequences of actions, through experimentation with the environment, in the two cases of immediate feedback and delayed feedback are described. The inferencing schemes allowing the robot to classify the information acquired from a reduced set of examples and to generalize its knowledge to a much wider problem-solving domain are also provided. A demonstration of the successful implementation of the algorithms on our HERMIES-IIB autonomous robot is then presented. 8 refs., 2 figs.

• Research Organization: Oak Ridge National Lab., TN (USA)
• DOE Contract Number: AC05-84OR21400
• OSTI ID: 6265108
• Report Number(s): CONF-890522-2; ON: DE89001525
• Country of Publication: United States
• Language: English

Related Subjects

99 GENERAL AND MISCELLANEOUS
990210* -- Supercomputers-- (1987-1989)
ALGORITHMS
ARTIFICIAL INTELLIGENCE
COMPUTERIZED CONTROL SYSTEMS
CONTROL SYSTEMS
DESIGN
EVALUATION
IMPLEMENTATION
LEARNING
MATHEMATICAL LOGIC
ROBOTS