GAIA: Making Virtual Space a Reality
- Paper number
IAC-06-A3.P.1.01
- Author
Dr. Cees Bil, Royal Melbourne Institute of Technology (RMIT), Australia
- Year
2006
- Abstract
When launched and in position in 2012, GAIA is set to measure properties of over 1 billion stars with an accuracy never before achieved. GAIA will measure over 1 billion stars in the Galaxy, observing each object up to 100 times during Gaia’s 5-year lifetime, providing position data of unprecedented accuracy (about 10 µas at 15th magnitude) as well as radial velocities and photometric measurements in 16 broad and medium band filters. The mission will provide a detailed database of the expanse of our galaxy and possible solar systems within it. GAIA is designed to follow the highly successful Hipparcos mission that resulted in the Hipparcos Tyco Catalogue of 118,000 stars.
The GAIA mission objective is to chart a three-dimensional map of our Galaxy and in the process reveal its composition, formation and evolution. GAIA will produce about 20 Terabytes of raw telemetry data that, after processing and reduction, will generate a database of the order of 1 Peta byte. This enormous amount of data pushes current computer technology and programming techniques to the limit. GAIA data is required to reside in a database in its entirety, due to the complex interaction of the algorithms that will operate on the data to derive distances, proper motions astrophysical properties and create the final three-dimensional model of the Galaxy.
Data mining or knowledge discovery is a relatively new area of database research. Data mining is defined as the nontrivial extraction of implicit, previously unknown, and potentially useful information from data. This is done by combining a variety of database, statistical and machine learning techniques. Different data mining algorithms have been developed to perform different types of analysis. Some algorithms search for repeating patterns or trends in a database. Others classify elements into groups based on their attribute values. For the data from GAIA, we are interested in data mining algorithms that perform classification and spatial location. These algorithms can be used to improve the efficiency of visualizing large, multidimensional datasets, which then can be presented in 3D virtual reality. Although GAIA will not launch before 2010 this work facilitates an insight into the data processing techniques required to create a useful tool for astrophysicists.
The paper will present a data mining technique developed by RMIT University and ESA proposed for retrieving information from the GAIA database. This tool allows astrophysicists to simulate planet formation, stellar system collisions, and origins of our galaxy.
- Abstract document
- Manuscript document
IAC-06-A3.P.1.01.pdf (🔒 authorized access only).
To get the manuscript, please contact IAF Secretariat.