“It was like being over the Bay Area in a plane,” she said. “There were the Oakland hills, there were the rivers, there was the sea. It looked like you could land and be at home. Except if you did, you’d be in trouble — it’s 180 degrees below zero there.”
Cabrol, a scientist at the SETI Institute in Mountain View, was inspired to dream up a robot that could be used to develop technology for the next probe dropped onto Titan — one she hopes lasts longer than Huygens, which sent data for about 90 minutes before petering out.
Specifically, Cabrol wanted to develop technology that could be used by a robot to explore Titan’s alien seas. She enlisted the help of the NASA Ames Intelligent Robotics Group, the agency’s top robotics and artificial intelligence research cluster, which is based on the Peninsula. SETI received a grant from NASA for the project.
Together, SETI and NASA scientists and engineers in the Bay Area conceived of a floating robot that could explore a lake on Earth, learning how to do science with relative autonomy. That technology could then be further refined for use on a robot sent to explore a sea on Titan.
The technology will help NASA develop a new generation of highly sophisticated interplanetary exploration robots, which can allow humankind to explore parts of our solar system that are far out of reach for human bodies. Further, scientists are beginning to explore the applications these developments could have on Earth as well.
The Planetary Lake Lander became a reality in 2011 after the team won millions in funding from NASA’s Astrobiology Science and Technology for Exploring Planets program. To test the robot, they chose a glacial lake in the central Chilean Andes where glacial warming is having an unusually strong effect — an ideal environment for training a robot that must adapt quickly to change.
Since 2011, the Lake Lander has been exploring its environment mostly on its own, with periodic visits from Cabrol and her team, including one last month.
Meanwhile, NASA has considered sending a floating robot to Titan. The proposed Titan Mare Explorer narrowly lost a funding bid last year to a 2016 Mars lander mission, but it’s likely to be considered for a future mission.
If it succeeds, it will use much of the technology developed for the Lake Lander.
THE PERFECT STORM
Titan holds special interest for scientists because it is the only other body in our solar system known to have seas, rivers and lakes, as well as a developed atmosphere complete with storms. The catch is that all that liquid and vapor are hydrocarbons, not water. Titan’s seas are mostly methane, a liquid inhospitable to Earthly life. Nonetheless, scientists are fascinated with the distant moon’s similarities to Earth and are eager to explore it.
Such a mission would come with major logistical hurdles, not the least of which is the billion-mile distance between Earth and Titan; the trip took Huygens seven years. Because Titan’s environment is ever-changing and there’s a two-hour delay in communications, a robot there would have to solve problems on its own or notice interesting things without the benefit of human direction.
During last month’s visit to the Chilean Andes, engineers from the Intelligent Robotics Group worked on the robot’s brainpower and autonomy. NASA roboticist Liam Pedersen, who lives in Potrero Hill, said one aim of the group is to improve the speed at which robots can perform simple tasks.
He noted that recent Mars rovers are already much improved in that regard compared to their predecessors, but they still have a long way to go until they compete with humans in many tasks.
“Say there’s a rock over there that looks interesting, so you walk over there, pick it up and look at it,” Pedersen said. “That’s an operation that takes a geologist 10 seconds, but it used to take a rover three to four days.”
Today, the operation takes less than a day — an improvement, but still far from human efficiency, he said.
PROGRAMMING THE HUMAN TOUCH
Robots have humans beat on some other tasks, said hardware lead Susan Lee of NASA, who lives in Mountain View.
“There’s a lot of things that robots are better at than humans, tasks where humans would get tired and bored and make mistakes,” she said.
For a trip to Titan, a robot would need to analyze data from its environment to determine when something interesting was happening, then focus its instruments and cameras accordingly. Changes in the environment that would be obvious to a human mind must be programmed into a robot. And since the communication delay is so long, an opportunity to observe something interesting might not be noticed by operators on Earth until it’s too late.
“I just have this image of sending a Lake Lander to Titan and floating right by a stream — the first stream ever seen outside the Earth, and we’re going to float right by it,” Pedersen said.
To try to avoid that scenario, the team floated the Lake Lander near streams and collected data about turbidity, or clarity; pH; chemistry; temperature and other factors to try to teach the robot how to “see” a stream.
The team also taught the robot how to detect a storm; it now notifies the team by email every time a storm arrives on its lake.
EYES ON THE ENVIRONMENT
A secondary goal of the project is to learn about how climate change is impacting the Andes, where deglaciation is occurring at one of the highest rates on Earth.
Since the lake is fed by the rapidly declining Echuarren Glacier, the data the robot has collected over the years is being used by researchers to determine how climate change is affecting life there.
But while the Lake Lander’s science is providing a grim picture of how life is changing on Earth, it’s also providing hope for future exploration of other planets. For Cabrol, the idea of sending a rover to Titan is NASA’s most exciting prospective space mission since it landed on the moon.
“I don’t want to minimize the rovers on Mars — landing and operating them is a challenge and an important one — but Titan is totally different,” she says. “It’s like a new frontier.”