You Can Now Help In Training NASA’s Rovers – So They Can Explore Mars More Effectively

Members of the general public can now educate an artificial intelligence algorithm to acknowledge scientific options in photographs taken by NASA’s Perseverance rover.

Artificial intelligence, or AI, can significantly change how NASA’s space missions research the cosmos. However, since all machine learning algorithms require human training, a recent experiment asks public members to identify scientifically noteworthy features in imagery collected by NASA’s Mars rover Perseverance.

The AI4Mars project relies on a previous one that used data from NASA’s Curiosity rover last year. Participants classified almost 500,000 images in the early stages of the project, using a tool to indicate qualities such as sand and rock that rover drivers at NASA’s Jet Propulsion Laboratory while selecting trajectories on Mars. SPOC (Soil Property and Object Classification) was the end product, an algorithm that could accurately detect these attributes 98 percent of the time.

SPOC is in progress, with plans to take it to Mars on a future spacecraft capable of even greater autonomous driving than Perseverance’s AutoNav technology.

Perseverance imagery will improve SPOC more by expanding the types of identifying labels applied to Martian objects. Currently, AI4Mars assigns labels to detect finer details, allowing users to choose between float rocks (rock islands) and nodules (Water-formed BB-size balls of minerals cemented together).

The basic idea is to improve an algorithm to help a future Mars rover find needles in a stack of data delivered from the Red Planet. Perseverance, which has 19 cameras, sends hundreds of photographs to Earth daily so that scientists may look for certain geological traits. However, time is of the essence: after those images have travelled millions of kilometres from Mars to Earth, the team members only have a few hours to develop the next set of instructions to send to Perseverance based on what they observe in the images.

“It’s not possible for anyone scientist to look at all the downlinked images with scrutiny in such a short amount of time, every single day,” said Vivian Sun, a JPL scientist who coordinates Perseverance’s daily operations and consulted on the AI4Mars project. “It would save us time if there was an algorithm that could say, ‘I think I saw rock veins or nodules over here,’ and then the science team can look at those areas with more detail.”

SPOC requires a lot of validation from scientists to ensure precise labelling. Even if it improves, the algorithm isn’t intended to take the place of more complex analyses conducted by human experts.

According to Hiro Ono, the JPL AI researcher who spearheaded the development of AI4Mars, a suitable dataset is necessary for a successful algorithm. The more individual pieces of data a machine learning system have, the more it learns.

“Machine learning is very different from normal software,” Ono said. “This isn’t like making something from scratch. Think of it as starting with a new brain. More of the effort here is getting a good dataset to teach that brain and massaging the data so it will be better learned.”

Artificial intelligence researchers can use tens of thousands of photographs to train their algorithms. But, unfortunately, there was no equivalent data stored for the surface of Mars previous to the AI4Mars initiative. The team would be ecstatic if it had 20,000 or more images in its archive, each with its own set of features.

According to JPL’s Annie Didier, who worked on the Perseverance version of AI4Mars, the Mars-data store could serve numerous uses. “With this algorithm, the rover could automatically select science targets to drive to. It could also store a variety of images onboard the rover, then send back just images of specific features that scientists are interested in,” she said.

Scientists will most likely not have to wait long for the algorithm to be beneficial to them. Before being launched into orbit, the device could be used to scan NASA’s enormous public collection of Martian data, making it easier for researchers to spot surface features in the images.

According to Ono, the AI4Mars team must make their dataset publicly available to benefit the entire data science community.

“If someone outside JPL creates an algorithm that works better than ours using our dataset, that’s great, too,” he said. “It just makes it easier to make more discoveries.”

Perseverance’s mission on Mars has a significant purpose of astrobiology. The rover will study Mars’ geology and former climate to prepare for future human Mars exploration missions.  

Following NASA missions, spacecraft would be sent to Mars in partnership with ESA (European Space Agency) to collect these sealed samples from the surface and return them to Earth for further examination.

The Mars 2020 Perseverance mission is part of NASA’s Moon to Mars exploration strategy, including Artemis Moon missions to help prepare for human exploration of Mars. 

Source: JPL

Leave a Reply

Your email address will not be published. Required fields are marked *