Spacecraft are great explorers, but they can be frustrating pen pals.
The farther from home a probe ventures, the longer its dispatches take to reach eager humans on Earth and the terser such reports must be. That’s why computer scientists and planetary scientists are teaming up to develop an algorithm that could potentially identify the most intriguing data an icy moon explorer mission collects, sending those tidbits to receivers first.
“We’re in this golden age of space exploration, and we have hundreds and hundreds of gigabytes of data flooding back from across the solar system,” Ashley Davies, a planetary scientist at NASA’s Jet Propulsion Laboratory in California, told Space.com. “It’s not possible to return all the data that you ever collect.”
Hence the interest in an algorithm to negotiate what to report first. A team based at JPL is developing a potential system to do just that for individual instruments on NASA’s Europa Clipper mission.
That spacecraft is due to launch in the mid-2020s to explore Jupiter’s icy moon Europa. The moon is one of the most intriguing worlds in our solar system for scientists interested in understanding whether life exists beyond Earth. An ocean hidden below Europa’s icy shell could potentially host microbial life similar to that found near deep-sea vents on Earth, and Clipper could collect information that would give scientists a more detailed understanding of the moon.
Europa Clipper will carry nine different science instruments, and the algorithm team is already working with the teams behind three of them, with more partnerships under discussion. The instruments will look for features like warm patches in the icy shell and plumes of seawater bursting out into space.
The idea behind the algorithm project is that it should be possible to train the spacecraft to spot the most promising data it gathers, then bump that to the front of the communications queue. It wouldn’t change what Europa Clipper does next, but it could mean scientists wouldn’t need to be quite as patient.
There’s just one problem: space-ready machines are … not the typical hardware computer scientists use. The system on board Europa Clipper will be able to run at speeds of up to 200 megahertz. “For comparison, that’s about the equivalent of an early ’90s desktop PC, if you took that and put that into orbit,” Kiri Wagstaff, a computer scientist at JPL who is leading the project, told Space.com. “We don’t really encounter machines that limited today in our day-to-day life.”
The stunted processing capacity means any space-bound algorithm has to be lean, and extremely so. It’s one of the two key challenges Wagstaff and her colleagues are facing in developing a system for Europa Clipper: The researchers have to come up with ways of flagging key data for the limited processor to execute in ways that are simple and quick. “We get questions about, well, ‘Are you using deep learning to help these spacecraft make these decisions?'” Wagstaff said. “And the answer is emphatically no, it’s simply not possible.”
The team’s other main challenge is that, after all, Europa Clipper itself doesn’t exist yet and hasn’t produced any data yet. Wagstaff and her colleagues are basing their current work on data gathered by other spacecraft and on simulations of what Clipper’s data could look like, but it’s not the same.
Right now, the algorithm isn’t a formal part of Europa Clipper, and there’s no guarantee it will be used during the mission. First, it needs to pass a series of tests designed to make sure the algorithm is trim enough to work in flight. “That’s kind of a go, no-go point,” Wagstaff said. “If it can’t fit into the available resources, you can’t use it.”
Wagstaff and her colleagues are also checking whether the algorithm can withstand the harsh radiation environment around this frigid moon. The computer on the spacecraft will be sheathed to protect it from radiation, but some particles will still sneak through. Scientists need to know whether such hits can derail a calculation.
But if the team can get the algorithm right, it could make Europa Clipper a more powerful mission, encouraging scientists to gather too much data to send home in the allotted time, with the assurance that they’ll see the most intriguing of it.
That’s the first step in using such algorithms to actively shape outer solar system missions. The process has already begun at Mars on NASA’s Curiosity rover, which can use its laser spectrometer to analyze rocks that meet current science priorities without waiting for instructions from Earth. “It all happens without any human in the loop,” Wagstaff said. “The rover itself decides, ‘This looks like an interesting rock, I will sample that,'” in time that it would otherwise spend twiddling its thumbs. “We’re getting this bonus science basically for free.”
Examples like that make a tempting lure for planetary scientists with more distant targets. But even if an algorithm doesn’t fly on Clipper, there’s plenty for scientists to look forward to, Davies said.
“Whatever happens, the data that we’re going to get back from Europa Clipper is going to be of immeasurably higher quality than what was achievable from previous missions,” he said. “We’re going to see a lot of things that previous instruments simply couldn’t have detected.” They may just need a little more patience.
Editor’s note: This story has been updated to clarify that the Mars Curiosity instrument that can operate autonomously is its laser spectrometer, not its drill. Email Meghan Bartels at firstname.lastname@example.org or follow her @meghanbartels. Follow us on Twitter @Spacedotcom and on Facebook.
(Image credit: All About Space magazine)
Tags: Alien, all, Beyond, Board, Challenge, Challenges, Change, Computer, Data, Facebook, Flight, free, Hardware, Image, it, Jet, jet propulsion laboratory, JPL, Mars, News, or, PC, Run, Science, Series, Space, story, Thing, Up, Us