An Application of Artificial Intelligence to Automatic Telescopes

KeRh Swansonswanson@pt olemy, arc. nasa. govTel: 415-604-6016NASA Ames Research CenterMS: 269-2, Moffett Field, CA 94035-1000, U.S.A.//JMark Drummond and John Bresina{drur_ond, bresina}@pt olemy, arc. nasa. govSterling SoftwareNASA Ames Research CenterMS: 269-2, Moffett Field, CA 94035-1000, U.S.A.AbstractAutomatic Photoelectric Telescopes (APTs) allow an astronomer to be removed from thetelescope site in both time and space. APTs "execute _ an observation program (a set of obser-vation requests) expressed in an ASCII-based language (ATIS) and collect observation resultsexpressed in this same language. The observation program is currently constructed by a Prin-cipal Astronomer from the requests of multiple users; the execution is currently controlled bya simple heuristic dispatch scheduler. Research aimed at improving the use of APTs is beingcarried out by the Entropy Reduction Engine (ERE) project at NASA Ames. The overall goalof the ERE project is the study and construction of systems that integrate planning, schedul-ing, and control. This paper discusses the application of some ERE technical results to theimprovement of both the scheduling and the operation of APTs.This paper subsumes previous versions which appeared in1. Proceedings o/the Remote Observing Workshop, held in Tuscon, AZ in April 1992.2. Proceedings o/the Steerable Automatic Lunar Ultraviolet Telescope Ezplorer (SALUTE)Workshop, held at NASA Ames Research Center, Moffett Field, CA in May 1992.