Scientist turns to AI and seeks to find masses of floating planets

Exoplanet hunters have found thousands of planets, most orbiting near their host stars, but relatively few alien worlds have been detected that float freely across the galaxy as rogue planets, unrelated to any stars. Many astronomers believe these planets are more common than we think, but our planet-finding techniques have not been up to the task of locating them.

Most of the exoplanets discovered to date have been discovered because they produce slight dips in the observed light of their host stars as they pass through the star’s disk from our perspective. These events are called transits.

NASA’s Nancy Grace Roman Space Telescope will investigate to discover many more exoplanets using powerful techniques available to a wide-field telescope. The stars in our Milky Way galaxy are moving, and random alignments can help us find rogue planets. When a floating planet aligns precisely with a distant star, it can make the star glow. During such events, the planet’s gravity acts as a lens that briefly magnifies the light from the background star. Although Roman can find rogue planets using this technique, called a gravitational microlens, there is a downside: the distance to the lens planet is poorly known.

Goddard scientist Dr. Richard K. Barry is developing a mission concept called the Contemporaneous LEnsing Parallax and Autonomous TRansient Assay (CLEoPATRA) to exploit parallax effects to calculate these distances. Parallax is the apparent shift in the position of an object in the foreground as seen by observers in slightly different places. Our brain operates the slightly different views of our eyes so that we can see depth as well. 19th-century astronomers first established distances to nearby stars using the same effect, measuring how their positions moved relative to background stars in photographs taken when Earth was on opposite sides. of its orbit.

It works a little differently with the microlens, where the apparent alignment of the planet and the distant background star is highly dependent on the position of the observer. In this case, two well-separated observers, each equipped with an accurate clock, would witness the same microlens event at slightly different times. The delay between the two detections allows scientists to determine the distance from the planet.

To maximize the parallax effect, CLEoPATRA would participate in a mission to Mars which launches around the same time as Roman, currently scheduled for the end of 2025. This would place it in its own orbit around the Sun which would reach a sufficient distance from Earth to effectively measure the parallax signal from the microlenses and fill in this missing information.

The CLEoPATRA concept would also support the PRIme-focus Infrared Microlensing Experiment (PRIME), a ground-based telescope currently equipped with a camera using four detectors developed by the Roman mission. The mass estimates of microlens planets detected by Roman and PRIME will be greatly improved by the simultaneous parallax observations provided by CLEoPATRA.

“CLEoPATRA would be a great distance from the main observatory, either Roman or a telescope on Earth,” Barry said. “The parallax signal should then allow us to calculate fairly precise masses for these objects, thus increasing scientific yield.”

Stela Ishitani Silva, a research assistant at Goddard and a Ph.D. student at the Catholic University of America in Washington, said understanding these floating planets would help fill some of the gaps in our knowledge about how planets were formed.

“We want to find several floating planets and try to get information about their masses, so that we can understand what is common or not common at all,” Ishitani Silva said. “Obtaining mass is important to understanding their planetary development.”

In order to efficiently find these planets, CLEoPATRA, which completed a mission planning lab study at Wallops Flight Facility in early August, will use artificial intelligence. Dr Greg Olmschenk, a postdoctoral researcher working with Barry, developed an AI called RAPid Machine learnEd Triage (RAMjET) for the mission.

“I work with certain types of artificial intelligence called neural networks,” Olmschenk said. “It’s a type of artificial intelligence that will learn through examples. So you give it a bunch of examples of what you want to find and what you want it to filter, and then it learns to recognize them. patterns in that data to try to find the things you want to keep and the things you want to throw away. “

Eventually, the AI ​​learns what it needs to identify and will only return important information. By filtering this information, RAMjET will help CLEoPATRA to overcome an extremely limited data transmission rate. CLEoPATRA will have to observe millions of stars every hour or so, and there is no way to send all this data to Earth. Therefore, the spacecraft will need to analyze the data on board and return only measurements from sources it detects as microlens events.

“CLEoPATRA will allow us to estimate many high precision masses for the new planets detected by Roman and PRIME,” Barry said. “And that will perhaps allow us to capture or estimate the actual mass of a free-floating planet for the first time, which has never been done before. So cool and so exciting. Really, it’s a new golden age for astronomy right now, and I’m just really excited about it. ”