Scientist looks to AI, lensing to find masses of free-floating planets

This illustration shows a Jupiter-like, solitary planet floating in space alone without any parent stars. Scientists from CLEoPATRA hope to increase the mass estimates for such planets through microlensing. Credit: NASA's Goddard Space Flight Center Conceptual Image Lab
Many thousands of planets have been discovered by exoplanet researchers. Most are located close to their host star, but very few have been found that orbit freely in the galaxy. These so-called "rogue planets" are not bound to any star and float freely throughout the galaxy. These planets are believed to be more common than we think, however our planet-finding methods have not been able to locate them.

The majority of exoplanets found to date were discovered because they cause slight dips in the observed starlight as they cross the star's disc from our perspective. These are known as transits.

NASA's Nancy Grace Roman Space Telescope (NASA) will perform a survey to find more exoplanets by using powerful techniques that are available to wide-field telescopes. Chance alignments are a way to find rogue planets. The Milky Way galaxy stars move. A star can brighten when a planet that is free to move aligns with a distant star. These events are caused by the planet's gravity acting as a lens to briefly magnify the background star's light. This technique, known as gravitational microlensing (or Roman), may help Roman find rogue stars. However, it has one problem: the distance to the lensing planet's location is not well understood.

Dr. Richard K. Barry, a Goddard scientist, is currently developing a mission concept called ContemporaneousLEnsing PARallax and Autonomous TRansient ASsay (CLEoPATRA). This mission concept will exploit parallax effects in order to calculate these distances. Parallax refers to the apparent shift in the foreground object's position as seen from slightly different locations. Because our brains have slightly different views, we are able to see depth with them. The same effect was used by 19th-century astronomers to determine the distances between nearby stars. They measured how the positions of background stars changed relative to Earth's orbit using photographs taken at opposite ends of the planet.

Microlensing is a different story. The apparent alignment of the distant background star and the planet depends greatly on the position of the observer. Two observers with precise clocks, separated by a few seconds, could witness the same microlensing event, but at slightly different times. Scientists can determine the distance of the planet by using the time delay between two detections.

To maximize the parallax effect CLEoPATRA would ride along a Mars-bound mission. It launches at the same time that Roman, which is currently scheduled for late 2025. This would put it in a orbit around the Sun, which would allow it to measure the microlensing paraallax signal from Earth and fill in the missing information.

The CLEoPATRA idea would also support PRIME-focus Infrared Microlensing Experiment. (PRIME) is a ground-based telescope that currently has a camera and four detectors built by the Roman mission. CLEoPATRA will provide simultaneous parallax observations that will significantly improve mass estimates of microlensing worlds as detected by PRIME and Roman.

This animation shows the concept of gravitational microscopicsing with a "rogue" planet, a planet that doesn't orbit a star. The warped space-time around a foreground planet causes light rays from the background star to bend when the rogue planet passes close to it. Credit: NASA's Goddard Space Flight Center/CI Lab

Barry stated that "CLEoPATRA would lie at a great far distance from the principal observatory," either Roman or a telescope in Earth. "The parallax signal will allow us to calculate very precise masses for these objects, increasing our scientific return."

Stela Ishitani Silva is a Goddard research assistant and a Ph.D student at the Catholic University of America, Washington. She believes that understanding the free-floating planets will fill in some of our gaps in understanding how they form.

Ishitani Silva stated that "we want to find multiple free-floating worlds and attempt to get information about their masses so we can understand what is similar or different." It is crucial to understand their planetary evolution that we can obtain the mass.

CLEoPATRA completed a Mission Planning Laboratory at Wallops Flight Facility in August and will now use artificial intelligence to find these planets. Barry's postdoctoral researcher, Dr. Greg Olmschenk has created an AI called RApid Machine learnEd Triage, (RAMjET), for the mission.

Olmschenk stated that he works with "certain types of artificial intelligence known as neural networks." It's an artificial intelligence type that learns from examples. You give it examples of what you want it to find and filter out the rest. Then it learns how to recognize patterns and find the items you want to keep or to discard.

The AI eventually learns what it is looking for and will only return important information. RAMjET will filter this information and help CLEoPATRA overcome a very limited data transmission speed. CLEoPATRA must watch millions of stars per hour, and there is no way to transmit all that data to Earth. The spacecraft will need to analyze all the data it receives and only send back measurements from sources it deems to be microlensing.

Barry stated that CLEoPATRA would allow us to calculate high-precision masses of new planets discovered by Roman and PRIME. It may also allow us to estimate or capture the mass of a free-floating world for the first time, something that has never been done before. It's so cool and exciting. It's really a new golden age in astronomy right at the moment, and I'm just so excited about it."

Explore more Rogue planets may outnumber stars