Rovers exploring planetary surfaces like Mars or the moon can encounter soft, loose and steep soils, the perfect ingredients for getting stuck.
But the "Mini Rover," a scaled down 3D-printed model of a NASA rover concept, built and tested by researchers at the Georgia Institute of Technology, has a gait that allows it to crawl and successfully conquer tricky terrain.
The findings were published in a study Wednesday in the journal Science Robotics.
Its origins date back to 2009, when NASA's Spirit rover became stuck in soft soil on Mars and couldn't be moved. It was used as a stationary science platform for three months until the rover stopped communicating in March 2010. The mission was officially concluded by NASA in 2011.
The soft granular soil, which is known as regolith, acts almost like a fluid when it's been disturbed on Mars, meaning it's easier to get stuck in it. (The moon has similar soil.)
After the loss of Spirit, engineers at NASA's Johnson Space Center in Houston built the Resource Prospector mission rover, or RP15 rover.
The concept involved a rover that allowed for wheel spinning combined with a lifting and sweeping motion by the rover's legs. But the mission itself was canceled in 2018.
Supported by the NASA National Robotics Initiative and the Army Research Office, researchers at Georgia Tech proposed building and testing a miniature version of the rover made from 3D-printed plastic. They dubbed this small-scale RP15 the Mini Rover.
"Our broad design goal was to replicate RP-15 drive functionality on a scaled-down robot that we could operate in our instrumented test bed," said Siddharth Shrivastava, study author and an undergraduate student in Georgia Tech's George W. Woodruff School of Mechanical Engineering, in an email to CNN.
Although the robot isn't an exact replica of the larger rover, which has motors encapsulated in its wheel diameter and a larger power-to-weight ratio, the Georgia Tech team was able to replicate the lifting, sweeping and wheel-spin motions. These motions were key for their tests. Their model, built using in-house 3D printers, has four wheeled legs that are powered by 12 motors.
Once the robot was assembled at Georgia Tech, it was time to test the Mini Rover on flat and sloped beds of poppy seeds. While they aren't exactly an analog of the surfaces these rovers might encounter on other planets, the 1 millimeter seeds have a number of features that make them suitable for experiments.
"Poppy seeds are quite a useful granular media for laboratory experiments with robots," said Andras Karsai, study coauthor and Georgia Tech School of Physics doctoral candidate, in an email to CNN.
"They are soft enough to crush safely between moving robotic joints if stuck, and do not produce any harmful dust."
The researchers started by putting Mini Rover through the same gait motion tests that RP15 experienced at Johnson Space Center and also explored ranges of motion that only the smaller rover could perform due to its diminutive size.
If only the wheels were activated, the robot would slip over the poppy seeds and sink down. When the sweeping motion of the robot's legs was added, the Mini Rover was able to use this unique locomotion method to paddle out of the poppy seeds as they shifted unpredictably. This prevented the rover from sinking more by creating seed mounds to its advantage that it could climb over.
This wheeled, sweeping and lifting motion created a successful crawling motion for the rover, even when one of its limbs was intentionally disabled. This crawl was dubbed "rear rotator pedaling" by the researchers.
"When loose materials flow, that can create problems for robots moving across it," said Dan Goldman, study coauthor and Dunn Family Professor in the School of Physics at the Georgia Institute of Technology, in a statement.
"This rover has enough degrees of freedom that it can get out of jams pretty effectively. By avalanching materials from the front wheels, it creates a localized fluid hill for the back wheels that is not as steep as the real slope."
Disabling lift, sweep or wheel spin individually hindered the rover. But the combination of all three created the most efficiency and reduced drag.
"This was an elegant experiment with clear results: Combining all three systems was necessary," Shrivastava said.
The same was true when they tilted the seed bed to simulate the rover climbing steep planetary hills.
The rover could successfully climb the hill when its front wheels stirred up the poppy seeds and pushed them back to its rear wheels, and the rear wheels wiggled in a side-to-side paddling motion. This created a smaller slope for the rear wheels to climb.
"Disturbing the granular media is actually key to the robot's performance," Karsai said. "By constantly agitating the granular media with its wheels, the robot forms a localized frictional fluid which it can effectively 'crawl' through.
"On sloped surfaces, this agitation also helps the media avalanche into local mounds that the rover robot can push off of. In brief, the 'robophysics' approach led to many surprising outcomes, which we could not have predicted through modeling alone."
Overall, the rover was able to continue up the hill. But a robot with only spinning wheel motion would have become stuck.
These tests were also carried out at Johnson Space Center with the full-size RP15 rover driving over a flat, moist sand bed, and they achieved the same results that were found with Mini Rover.
Going forward, the physicists and engineers want to test the larger RP15 and Mini Rover on steeper slopes using different soils and terrains to test more functionality. They also want to add local terrain sensing, which could help the rover detect its surroundings. Shrivastava is also curious to explore downhill locomotion.
The capabilities demonstrated in the study could be used to inform the design of next-generation rovers that could visit previously unexplored areas like the lunar poles, or even challenging spots on Earth, the researchers said.
"The work reported in this paper supports the design of rovers to access steep and soft soils, where scientifically interesting samples and materials might exist," Robert Ambrose, study coauthor and division chief of the software, robotics and simulation division at NASA's Johnson Space Center.
"For lunar exploration, that includes driving into dark craters and regions that are permanently shadowed at the Moon's poles. For Mars it might mean operating in soft sands and other deposits that are treacherous for rovers."
"Personally, I'm eager to see these strategies adapt beyond the laboratory within engineering groups developing locomotion platforms for the moon, Mars and beyond," Shrivastava said. "Aside from our species' launch vehicles, I believe robust extraterrestrial locomotion will serve as the bedrock for future human settlement beyond our planet."