In 2023, NASA plans to launch the mission of the Europa Clipper a robotic explorer to study Jupiter's mysterious moon. The purpose of this mission is to explore the ice cover and the interior of Europe to learn more about the composition, geology and interaction of the moon between the surface and the underground. Most of all, the purpose of this mission is to shed light on whether or not life can exist in Europe's inland ocean.
This presents many challenges, many of which stem from the fact that Europa Clipper will be very far from Earth when conducting its scientific operations. To cope with this, a team of researchers from the NASA Jet Propulsion Laboratory (JPL) and Arizona State University (ASU) have developed a series of machine learning algorithms that will allow the mission to explore Europe with a degree of autonomy.
How These Algorithms Can Assist in Future Deep Space Exploration Missions were the subject of a presentation held last week (August 7) at the 25th ACM SIGKDD Knowledge Discovery and Extraction Conference in Anchorage, Alaska. This annual conference brings together researchers and practitioners in data science, data mining and analytics from around the globe to discuss the latest developments and applications.
When it comes to it, communication with missions in deep space takes a long time. , challenging work. When communicating with missions on the surface of Mars or in orbit, it can take up to 25 minutes to reach them from Earth (or back again). On the other hand, sending signals to Jupiter can take anywhere from 30 minutes to an hour, depending on where it is orbiting Earth.
As the authors point out in their study, spacecraft activities are usually transmitted in a pre-planned scenario rather than in real-time commands. This approach is very effective when the position, environment and other factors affecting the spacecraft are known or can be predicted in advance. However, it also means that mission controllers cannot respond to unexpected real-time developments.
As Dr. Kiri L. Wagstaff, Principal Investigator at NASA's JPL Machine Learning and Instrumental Autonomy Group, explained to the Universe today via email:
"Exploring a world too far to allow direct human control is a challenge. All activities must be pre-written. The rapid response to new discoveries or changes in the environment requires the spacecraft itself to make the decisions we call spacecraft autonomy. In addition, operating a billion miles from Earth means that data rates are very low.
" The spacecraft's ability to collect data exceeds what can be sent back. This raises the question of what data should be collected and how priority should be given. Finally, in the case of Europe, the spacecraft will also be bombarded with intense radiation, which could damage the data and cause the computer to reset. Tackling these dangers also requires autonomous decision making. "
For this reason, Dr. Wagstaff and her colleagues began to look for possible methods of analyzing on-board data that would work anywhere and whenever direct human oversight was not possible. These methods are especially important in dealing with rare, transient events whose occurrence, location, and duration cannot be predicted.
This includes phenomena such as dust devils that have been observed on Mars, meteoric effects, lightning strikes on Saturn, and ice jets emitted by Enceladus and other bodies. To cope with this, Dr. Wagstaff and her team looked at the latest advancements in machine learning algorithms that allow for a degree of automation and independent decision making in calculations. As Dr. Wagstaff put it:
"Machine learning methods allow the spacecraft itself to examine the data when it is collected. The spacecraft can then identify which observations contain interesting events. This may influence the prioritization of the downlink. The goal is to increase the chance of the most interesting discoveries being erased first. When data collection exceeds what can be transmitted, the spacecraft itself can dig up additional data for valuable scientific elements.
"The analysis of the ship may also allow the spacecraft to decide which data to collect based on what it has already discovered. This was demonstrated in Earth's orbit using the autonomous science craft and the Mars surface experiment using the Mars Science Laboratory (Curiosity) AEGIS system. Autonomous, responsive data collection can greatly accelerate research. We seek to extend this capacity to the external solar system. "
These algorithms have been specifically designed to support three types of scientific research that will be critical to the mission of the Europa Clipper . These include the detection of thermal anomalies (hot spots), composite anomalies (unusual surface minerals or deposits), and active jets of ice from Europe's ocean floor.
"In this setting, the calculations are very limited," says Dr. Wagstaff. "The spacecraft computer operates at a speed similar to a desktop computer from the mid to late 1990s (~ 200 MHz). Therefore, we prioritize simple, efficient algorithms. A side benefit is that the algorithms are easy to understand, implement, and interpret. "
To test its method, the team uses its algorithms for both simulated data and observations from space missions to moons of Jupiter and other planets in the solar system. These include the spacecraft Galileo which carried out spectral observations of Europa to determine its composition; the spacecraft Mars Odyssey seeking the thermal anomalies of Mars; and Hubble Space Telescope observations on loop activity in Europe.
The results of these tests show that each of the three algorithms demonstrates a sufficiently high efficiency to meet the scientific goals outlined in the Planetary Decade of Science for 2011. They include "confirmation of the presence of an inland ocean characterizing the ice shell of the satellite. and enabling understanding of its geological history 'by Europa to confirm' the potential of the external solar system as a habitat for life '.
In addition, these algorithms could have far-reaching implications for other robotic missions to deep space destinations. Beyond the system of moons in Europe and Jupiter, NASA hopes to explore the moon moons of Entelud and Titan of Saturn for possible signs of life in the near future, as well as destinations that are further away (such as the moon of Neptune, Triton, and even Pluto). But the apps don't stop there. As Dr. Wagstaff says:
"The autonomy of spacecraft allows us to investigate where humans cannot go. This includes distant destinations such as Jupiter and places beyond our own solar system. It also includes closer environments that are hazardous to humans, such as the seafloor bottom or high radiation settings here on Earth. "
It's not hard to imagine a near future where semi-autonomous robotic missions are able to explore. external and internal achievements of the solar system without regular human supervision. Looking further into the future, it's not hard to imagine an era where fully autonomous robots are capable of exploring beyond solar planets and sending their discoveries home.
In the meantime, the semi-autonomous Europa Clipper can find the evidence we are all waiting for! These would be biosignatures that prove that there is indeed life beyond Earth!
Further reading: KDD 2019, Study (PDF)