Simulating extraterrestrial environments on Earth has always been a challenge. Our planet has a pleasant atmosphere, reasonable temperatures, and a moderate amount of gravity, unlike the rest of the solar system. Or maybe that’s just because we think that way because we adapted to how it is here as we evolved here. In either case, the physical environment here makes it difficult for us to set up test environments that can accurately test probes going to other parts of the solar system. Many times, it involves vacuum chambers, air conditioners and heaters pumping hot and cold air into them, and soil simulant – lots and lots of soil simulant. But, according to a new paper from researchers at the University of Wisconsin-Madison, we’ve been neglecting one important aspect of these tests, and it might be the reason Spirit eventually got permanently stuck on Mars – sand is affected by gravity too.
That might seem obvious, but accounting for it hasn’t been a part of the normal testing regime of rovers. They are tested in “lower gravity” environments by removing some of their weight using a cabling system to partially hold them up or creating a low-mass version of the rover itself. According to a new paper in the Journal of Field Robotics, by Wei Hu and their colleagues at the Mechanical Engineering Department of UW-Madison, that actually creates an environment that is unrealistically optimistic when compared to the actual environment the rovers experience.
Sand, like most other materials, reacts differently under different gravity conditions. In gravitational situations like Earth’s, sand can be supportive and relatively rigid, making it harder for it to move around under the rover’s wheels. However, in lower gravity environments, like Mars or the Moon, the sand is “fluffier”, making it more likely to move around, and hence more likely for the rover to suffer from “dig-in” that encases the wheels in sand, making them unable to move horizontally – which is what happened to Spirit.
Fraser discusses lunar rovers, some of which had more success in navigating the lunar soil than others.
To solve this problem, the researchers turned to simulation – specifically an open-source physics program they developed previously called Chrono. Testing their theory of how sand operates differently required them to model the Volatiles Investigating Polar Exploration Rover (VIPER), originally intended to go to the Moon relatively soon. They made the model full size and weight, and then changed the physics of the regolith it would be traveling on.
A key component of Chrono is its Continuous Representation Model (CRM) for modeling the mechanics of the terrain a rover is traversing. CRM uses a technique akin to fluid dynamics, called Smoothed Particle Hydrodynamics (SPH), to model how sand interacts with each other, though there’s some debate in the community over whether lunar and Martian regolith can be considered “smoothed”.
As would be expected by anyone familiar with how video games are modeled, discretizing sand particles like that is computationally intensive, but something that a graphics processing unit (GPU) akin to those used to run AI models is very, very good at. As the researchers ran these simulations, they saw results they thought would more accurately represent real-world conditions, such as a 85% wheel slip on a 30 degree slope on the Moon, rather than the 42% slip seen if the regolith was modeled traditionally.
Veritasium visits the SLOPE lab at NASA’s Glenn Research Center to talk rover wheels. Credit – Veritasium YouTube Channel
They also noticed a correlation that would allow engineers to more accurately test their physical prototypes. Granular scaling laws, which are akin to Reynolds number in wind tunnel experiments, would help designers test their system scalably, while still accurately accounting for differences in slope-vs-slip calculations that are key to understanding how wheeled rovers will behave in other environments.
Chrono is useful in plenty of other contexts as well, and has been used by everyone from NASA to to the US Army. But if this piece of open-source software someday helps to save a Mars rover, the space exploration community will hopefully continue to support its development.
Learn More:
University of Wisconsin – Madison – Robotic space rovers keep getting stuck. UW engineers have figured out why
W. Hu et al – A Study Demonstrating That Using Gravitational Offset to Prepare Extraterrestrial Mobility Missions Is Misleading
UT – End of the Road for Spirit Rover
UT – Perseverance Has Been Carrying a Rock in its Wheel for Over 100 Days