Concept formation and generalization based on experimentation by an autonomous mobile robot
P.F. Spelt, G. deSaussure, Eric Lyness, Gabriel Oliver, M. Silliman · OSTI OAI (U.S. Department of Energy Office of Scientific and Technical Information) · 1989
The Center for Engineering Systems Advanced Research (CESAR) conducts basic research in the area of intelligent machines. In this paper, we describe our approach to a class of machine learning problems which involves autonomous concept formation using feedback from trial-and-error learning. Our formulation was experimentally validated on an autonomous mobile robot, which learned the task of control panel monitoring and manipulation for effective process control. Conclusions are drawn concerning the applicability of the system to a more general class of learning problems, and implications for the use of autonomous mobile robots in hostile and unknown environments are discussed. 9 refs., 5 figs.