A Blog by

Cave-Exploring Snake Robot Gets Inspiration From Sidewinders

Three years ago, a robotic snake called Elizabeth slithered into Egyptian caves to search for long-hidden ships.

The caves lie on Egypt’s east coast, and contained the dismantled remnants of vessels that the Egyptians used to sail the Red Sea. They were discovered about a decade ago and some have surrendered their secrets with relative ease. Others, however, are too dangerous and unstable for people to explore.

Enter Elizabeth. The serpentine robot, built by Howie Choset at Carnegie Mellon University and named after his wife*, was designed to explore spaces that humans cannot. She can slide over rough terrain, slink through tight cracks, and manoeuvre around rubble. During her Egyptian field test, she performed beautifully, with one major exception: When the team tried to drive her up sandy slopes, she slipped and slid.

Real snakes face the same problem, and many desert-dwelling species have solved it through a bizarre technique called sidewinding. It’s a very counter-intuitive style of movement. From above, it looks like the snake is travelling sideways in a beautiful undulating wave. But it leaves behind a series of straight tracks, each the length of its body.

The trick to understanding the technique is to realise that the snake is never sliding. Instead, it is constantly picking itself off from its current position and laying itself down in a new spot. The head goes first, and the rest of the body follows. But before the body catches up completely, the head is off again. At any point of time, the snake is only touching the ground with two short parts of its body. That’s why it moves in a wave, but leaves a straight track.

Sidewinding is perfect for negotiating dunes. Rather than pushing against slippery sand, the snake’s rolling motion means that it’s mostly in static contact with its surface. Many species of snake can do this, but only two have truly mastered the technique—a rattlesnake from the US and Mexico, and a horned viper from Angola and Namibia. Confusingly, both are called sidewinders.

Choset’s robot Elizabeth could sidewind, but not very well. It was missing something that its real counterparts were doing.

Elizabeth, the robot snake. Credit: Nico Zevalios and Chaohui Gong.
Elizabeth, the robot snake. Credit: Nico Zevalios and Chaohui Gong.

To discover that mystery ingredient, Choset teamed up with Daniel Goldman from the Georgia Institute of Technology. For decades, Goldman has been fascinated by how animals move on and through sand. He has studied baby sea turtles as they clamber over a beach, and a pointy-nosed lizard called the sandfish as it swims through sand. And his team have built robots that emulate these animals, to reveal the physics behind their movements. Guy knows sand; guy knows robots. And as luck would have it, he was already starting to study sidewinders.

The team, led by postdoc Hamidreza Marvi and student Chaohui Gong, worked with six sidewinders (the American kind) from Zoo Atlanta. They put the snakes on a sandy trackway that could be inclined at different angles. They even trucked in sand from Arizona’s Yuma Desert to give the snakes material that they would normally face in the wild. “They’re excellent study subjects,” says Goldman. “They sidewind on command. Put them in a container and off they go.”

At first, the team assumed that as the track got steeper, the sidewinders would respond by digging their bodies more firmly into the ground, just like we would if we climbed a steep dune. They didn’t. Instead, they kept more of their body in contact with the ground, giving themselves more purchase on increasingly treacherous slopes. As the researchers raised the flat track to a 20 degree incline, the sidewinders compensated by laying down twice as much body.

The team also tested 13 other species of rattlesnake from Zoo Atlanta. None of them sidewind naturally, and none of them could negotiate the same slopes that the sidewinders could. They tried to climb straight up, and failed. “It was quite amusing,” says Goldman. “One of the comments we got from our reviewers was that it was obvious what the sidewinders do. Well, it wasn’t obvious to the other snakes!”

The team then programmed Elizabeth to mimic the sidewinders, and found that she suddenly became much better at moving up slopes. Her performance revealed that snakes have to stick within a certain range of contact lengths, and this range narrows as the slopes get steeper. If they don’t lay down enough body, they slip. If they lay down too much, they can’t lift the rest of themselves effectively, and run into the sand in front of them. They end up digging a hole, rather than making progress.

So, by playing with their robot, the team understood more about what the snakes do. And by studying the snakes, they improved their robot. “Using our understanding of fundamental engineering, we advanced these robots very far but we couldn’t get them up sandy hills,” says Choset. They only surmounted that final hurdle by studying nature.

Choset thinks that the snake-bots have many possible uses. They could search for survivors trapped in collapsed buildings. They could also inspect dangerous environments like nuclear storage facilities. And, of course, they could explore archaeological sites. “If we have the opportunity to return to Egypt, we’d use this capability,” he says. “Archaeology is like search and rescue except everyone’s been dead for thousands of years so there’s no rush.”

* Choset tells me that there was a second snake robot called Howard, but he was lost in some airline baggage mix-up. Samuel L. Jackson was unavailable for comment.

Reference: Marvi, Gong, Gravish, Astley, Travers, Hatton, Mendelson, Choset, Hu & Goldman. 2014. Sidewinding with minimal slip: Snake and robot ascent of sandy slopes. Science http://dx.doi.org/10.1126/science.1255718

More on animal robots:

6 thoughts on “Cave-Exploring Snake Robot Gets Inspiration From Sidewinders

  1. Intriguing. One statement is incorrect; there are not just two species that sidewind regularly … The genera Cerastes, Pseudocerastes, and Echis do this as well as the two you mention. Infact, species of the genus Echis occur throughout Africa and southwestern Asia.

  2. Sidewinding has also been elicited experimentally in a variety of smallish woodland snakes that may (this is unconfirmed) use it in the wild as escape locomotion on loose leaf-litter substrates (statement of interest: I published that), and was also recently documented as escape behaviour in a 9-foot Green Anaconda. Sometimes it’s just the fastest way to get around.

  3. I can personally add that Howie Choset is a great collaborator and great fun to spend time with. I had the great fortune of working with him a bit in Pittsburgh when I was faculty at Chatham University just down the road from CMU. Top notch scientist and a great individual.

Leave a Reply

Your email address will not be published. Required fields are marked *