Visit ASIAA Homepage UM2010 at CFHT
2010 CFHT Users Meeting
November 16-18, 2010
ASIAA Auditorium, Taipei, Taiwan

Poster Presentation

Automated Service Observing (ASO) - realtime scheduling of astronomical observations using Artificial Intelligence

Author(s): B. Mahoney, K. Thanjavur

Presenter: Karun Thanjavur (CFHT)

Automated Service Observing (ASO) is a natural evolutionary step from Queued Service Observing (QSO), which replaced classical observing to become the de facto observing paradigm at CFHT over the past decade. In QSO, each morning the queue coordinator (QC) prepares several queues, each tuned for a particular set of expected sky conditions (IQ, sky brightness, etc) and known queue constraints (time restricted targets, agency balance, etc). At night, based on the evaluation of possibly varying sky conditions, the observer decides on the optimal queue to implement, and when to change between queues. ASO aims to supplement, and eventually replace, this human decision making and scheduling process, based on fuzzy (human) logic, with well tested tools drawn from Artificial Intelligence (AI). Here, we present preliminary results from our tests at CFHT using two AI algorithms, Markov Decision Processes (MDP), and Genetic Algorithms (GA), for realtime scheduling of observations. Our initiatives will closely parallel and greatly augment the value of the Observatory Automation Project (OAP) which is in its final stages of implementation to permit remote observing at CFHT.

ASIAA will not contact participants for credit card information by phone. Privacy and Security Policy