NASA is currently studying concepts for flagship missions (habex and luvoir) with the goal of directly imaging a sample of earth-like planets in the habitable zones of nearby sun-like STARS. A…

This federal contract award from NASA is for a grant to the Massachusetts Institute of Technology (MIT) to develop more reliable catalogs of exoplanet candidates from the Kepler mission data. The goal is to produce the most sensitive catalog of Kepler planet candidates to date, which will enable more accurate measurements of the occurrence rate of Earth-like planets in the habitable zones of Sun-like stars. This is a critical piece of information for designing future flagship missions aimed at directly imaging Earth-like exoplanets. The $475,006 grant has a period of performance through July 2026. There is no set-aside designation for this award, indicating MIT's broad capabilities across federal scientific research contracts and grants.

NASA is currently studying concepts for flagship missions (habex and luvoir) with the goal of directly imaging a sample of earth-like planets in the habitable zones of nearby sun-like STARS. A critical ingredient to these mission concept studies is the prevalence of such earth-like planets if earth-like planets are intrinsically rare, larger and more expensive telescopes are required to conduct the survey. Presently, the best constraints on the occurrence of earth-sized habitable-zone planets around sun-like STARS come from the kepler mission, but the kepler planet-candidate catalog is incomplete and unreliable in this regime. As a result, published estimates of the occurrence rate of earth analogs, determined by extrapolating occurrence rates from larger and hotter planets, vary by over an order of magnitude. This level of uncertainty is an obstacle in the design of future flagship missions. We propose to apply state-of-the-art machine learning techniques to the search for exoplanets to address this problem. Our objective is to make the first measurement of the occurrence rates for earth analogs that does not rely on extrapolations from larger and hotter planets. Key to this objective will be our work to produce a new, more reliable catalog of kepler planet candidates with higher sensitivity to earth analogs. We will construct this catalog by performing a highly sensitive search for planet candidates and using modern machine learning techniques to reliably select the true planet signals from the false alarms. We have recently demonstrated this strategy can indeed uncover previously missed exoplanets. Increasing sensitivity to earth analogs in the kepler dataset, without sacrificing the reliability of the planet candidate catalog will enable more accurate occurrence rates. The result of our program will be the most sensitive catalog of kepler planet candidates produced to date. We will leverage this new data product to produce an accurate measurement of the occurrence of earth analogs. Our program will support the salary of a graduate student and summer salary for the PI. Our proposed work is relevant to the astrophysics data analysis program because IT is focused on analysis of archival data from the kepler mission.