The best life-hunting robot may look like a snake

Penetrating the ice on Enceladus may be a better strategy than drilling into it

Worlds with icy oceans, such as Europa or Enceladus, are among the most promising places to search for extraterrestrial life in the solar system because they have liquid water. But to determine if anything is hiding in their alien oceans, we need to overcome ice sheets tens of kilometers thick. Any robots we send out to make their way through the ice will have to do most of the work themselves, because communicating with these moons takes up to 155 minutes.

Researchers, working NASA's Jet Propulsion Laboratory project to develop technology called the Exobiology Extant Life Surveyor (EELS) may find a solution to both of these problems. It consists of using an artificially intelligent snake robot. And it has already been built.

Geysers on Enceladus

The most popular idea so far for breaking through the ice sheets of Enceladus or Europa has been thermal drilling, a technique used to study glaciers on Earth. In this case, a hot drill is used, which simply melts the ice. “Many people are working on different approaches to thermal drilling, but they all face the problem of sediment accumulation, which affects the amount of energy needed to move significantly through the ice sheet,” says Matthew Glinder, EELS project manager.

So instead of drilling new holes into the ice, the EELS team focused on using existing ones. The Cassini mission discovered geyser-like jets spewing water into space from holes in the ice sheet near Enceladus's south pole. “The concept was to land near the geyser vent, then the robot would crawl along the surface of the satellite, go down into the vent, explore it and then go down through the vent into the ocean,” says Matthew Robinson, EELS project manager.

The problem was that Cassini's best images of the area where the vehicle would land have a resolution of about 6 meters per pixel, which means that serious obstacles to the landing may go undetected. What's worse is that these close-up shots were monocular, which meant we couldn't properly identify the terrain. “Look at Mars. First we sent an orbiter. Then we sent a lander. Then we sent a small robot. And then we sent a big robot. This exploration paradigm allowed us to get very detailed information about the terrain,” says Rohan Thakker, project leader autonomous control of EELS. “But it takes seven to 11 years to get to Enceladus. If we followed the same paradigm, it would take us a century,” he adds.

All-terrain snakes

To cope with unfamiliar terrain, the EELS team created a robot that can go through almost anything – a versatile snake-like design about 4.4 meters long and 35 centimeters in diameter, inspired by biological examples. She weighs about 100 kilograms (at least on Earth). It consists of 10 almost identical segments. “Each of these segments uses a combination of a shape-changing actuator and an actuator that rotates screws mounted on the outside of the segments to propel the robot through its environment,” explains Glinder. Using these two types of actuators, the robot can move using what the team calls “skin traction,” which relies on the rotation of screws, or using one of various types of shape-based movements, which rely on shape actuators. “The robot is able to move simply by being pressed against a surface,” says Glinder.

The standard set of sensors is mounted on the head and includes a set of stereo cameras that provide a 360-degree viewing angle. There are also inertial measurement units (IMUs), which use gyroscopes to estimate the robot's position, and lidar sensors. In addition, the robot has a sense of touch. “We're going to put torque sensors in each segment. So we'll have direct torque plus direct sense of resistance at each joint,” Robinson explains. All this should allow the EELS robot to safely climb and descend the vents of Enceladus, stay in place during eruptions by hugging the walls, and even navigate only by touch if the cameras and lidar are not working.

But perhaps the most difficult part of creating the EELS robot was its brain.

Space Snake Brain

Manually controlling the EELS on Enceladus from Earth was out of the question due to huge communication delays, so the team went for almost complete autonomy. Ground control will be limited to issuing general commands such as “investigate this area” or “search for life.” “Imagine something like software that allows a Tesla to drive autonomously—only your car has 48 steering wheels, 48 ​​sets of pedals, and moves in a space where there are no roads, no stop signs, no speed limits,” explains Thakker. The AI ​​driving the EELS robot was built on a hierarchical, layered software architecture consisting of two categories of modules called evaluators and controllers.

The lowest level evaluators take information from internal sensors, such as the IMU and torque sensors in the segments, and use it to determine the robot's condition—whether it is falling, sliding, or hitting something. At a higher level are evaluators who build a map of the environment and determine the location of the robot based on data from cameras and lidars. The highest level evaluator considers risk and decides when to move quickly and when to move cautiously. The controllers responsible for taking action range from basic drive control systems at the lowest level to task and motion planning at the highest level.

“There are two sides to your mind. There is an intuitive side, which is fast, biased and very low power, and there is a logical side, which asks questions, evaluates the answers and tries to understand what is really going on. We try to use the same circuit where there is two such subsystems,” says Thakker. The intuitive part of EELS was built using machine learning, in which the robot taught itself how to move. The logical part is a model based on physics, with strictly fixed safety rules that do not allow the robot to exceed a certain speed, overcome slopes with a certain slope, and so on. All this should help the EELS robot get comfortable on foreign ice worlds.

If he gets there at all.

“We are not currently part of any flight mission,” Robinson says. However, he said the EELS architecture could be used in many other places, including Earth. “When we tested EELS on the Athabasca Glacier in Canada, we used it for real scientific research. We developed a scientific instrument that measured the salt content of the water flowing in the glacier. The robotic snake has terrestrial and space applications. It can be used for search -rescue efforts, studying the pile of debris, etc. But Enceladus remains a source of inspiration for us,” says Robinson.