Having a light touch can make a big difference in how well animals and robots move across surfaces such as snow, sand, and leaves.
Scientists studying how running lizards, geckos, crabs and a robot named Sandbot moved across a trackway filled with poppy seeds and glass spheres found that how far an animal’s legs or robot’s wheels penetrate into the surface changes performance.
Says Dan Goldman, professor of physics at the Georgia Institute of Technology:
“You need to know systematically how ground properties affect your performance with wheel shape or leg shape, so you can rationally predict how well your robot will be able to move on the surfaces where you have to travel. When the ground gets weak, certain animals seem to still be able to move around independently of the surface properties. We want to understand why.”
For years, researchers have been using trackways filled with granular material to study how animals and robots move, but in the past they had only used a fluidized bed to set the initial compaction of the material.
Graduate student Feifei Qian shows details of the Sandbot, a bio-inspired hexapedal robot. Credit: John Toon, Georgia Tech
In the new study they used variations in continuous air flow—introduced through the bottom of the device—to vary the substrate’s resistance to penetration by a leg or wheel. The researchers compare the trackway to wind tunnels used for aerodynamic studies.
“By varying the air flow, we can create ground that is very, very weak—so that you sink into it quite easily, like powdery snow, and we can have ground that is very strong, like sand. This gives us the ability to study the mechanism by which animals and robots either succeed or fail.”
Long Feet and Toes
Using a bio-inspired hexapedal robot known as Sandbot as a physical model, the researchers studied average forward speed as a factor of ground penetration resistance—the “stiffness” of the sand—and the frequency of leg movement.
The average speed of the robot declined as the increased air flow through the trackway made the surface weaker. Increasing the leg frequency makes the speed decrease more rapidly with increasing air flow.
Sandbot moves in a trackway filled with poppy seeds. Credit: John Toon, Georgia Tech
The five animals—with different body plans and appendage features—all did better than the robot, with the best performer being a lizard collected in a California desert. The speed of the C. draconoides wasn’t slowed at all as the surface became easier to penetrate, while other animals saw performance losses of between 20 and 50 percent.
“We think that this particular lizard is well suited to the variety of terrain because it has these ridiculously long feet and toes,” Goldman says.
“These feet and toes really enable it to maintain high performance and reduce its penetration into the surface over a wide range of substrate conditions. On the other hand, we see animals like ghost crabs that experience a tremendous loss of performance as the substrate changes, something that was surprising to us.”
The robot lost 70 percent of its speed even with wheels designed to lighten its pressure on the surface.
Zebra-tailed Lizard – Callisaurus draconoides. By Gregory “Slobirdr” Smith.
Skiers and beachcombers can certainly understand why. As the surface becomes easier for a ski or foot to penetrate, more energy is required to move and forward progress slows. Human and skiers haven’t evolved solutions to that problem, but desert-dwelling creatures have. The study will help researchers understand how they do it.
“The magic for us is how the animals are so good at this,” Goldman says. “There’s a clear practical application to this. If you can get the controls and morphology right, you could have a robot that could move anywhere, but you have to know what you are doing under different conditions.”
As part of the research, graduate students Feifei Qian and Tingnan Zhang used a terradynamics approach based on resistive force theory to perform numerical simulations of the robots and animals. They found that their model successfully predicted locomotor performance for low resistance granular states.
“This work expands the general applicability of our resistive force theory of terradynamics,” Goldman says. “The resistive force theory, which allows us to compute forces on limbs intruding into the ground, continues to work even in situations where we didn’t think it would work. It expands the applicability of terradynamics to even weaker states of material.”