Close

The Cheetah's Fluffy Tail Points the Way for Robots With High-Speed Agility

It’s not the mass, it’s the aerodynamics—and the result could be more sprightly machines

4 min read
CMU Minitaur with tail
Photo: CMU

Almost but not quite a decade ago, researchers from UC Berkeley equipped a little robotic car with an actuated metal rod with a weight on the end and used it to show how lizards use their tails to stabilize themselves while jumping through the air. That research inspired a whole bunch of other tailed mobile robots, including a couple of nifty ones from Amir Patel at the University of Cape Town.

The robotic tails that we’ve seen are generally actuated inertial tails: a moving mass that goes one way causes the robot that it’s attached to to go the other way. This is how lizard tails work, and it’s a totally fine way to do things. In fact, people generally figured that many if not most other animals that use their tails to improve their agility leverage this inertial principle, including (most famously) the cheetah. But at least as far as the cheetah was concerned, nobody had actually bothered to check, until Patel took the tails from a collection of ex-cheetahs and showed that in fact cheetah tails are almost entirely fluff. So if it’s not the mass of its tail that helps a cheetah chase down prey, then it must be the aerodynamics.

The internet is full of wisdom on cheetah tails, and most of it describes “heavy” tails that “act as a counterbalance” to the rest of the cheetah’s body. This makes intuitive sense, but it’s also quite wrong, as Amir Patel figured out:

The aerodynamics of cheetah tails are super important, and actually something I discovered by accident! Towards the end of my PhD I was invited to a cheetah autopsy at the National Zoological Gardens here in South Africa. The idea was to weigh and measure the inertia of the cheetah tail because no such data existed. Based on what I’d seen in wildlife documentaries (and speaking to any game ranger in South Africa), the cheetah tail is often considered to be heavy, and used as a counterweight.

However, once we removed the fur and skin from the tail during the autopsy, it was surprisingly skinny! We measured it (and the tails of another 6 cheetahs) as being only about 2 percent of the body mass—much lower than my own robotic tails. But the fur made up a significant volume of the tail. So, I figured that there must be something to it: maybe the fur was making the tail appear like a larger object aerodynamically, without the weight penalty of an inertial tail.

A few years ago, Patel started to characterize tail aerodynamics in partnership with Aaron Johnson’s lab at CMU, and that work has lead to a recent paper published in IEEE Transactions on Robotics, exploring how aerodynamic drag on a lightweight tail can help robots perform dynamic behaviors more successfully.

The specific tail design that Minitaur is sporting in the video above doesn’t look particularly cheetah-like, being made out of carbon fiber and polyethylene film rather than floof, and only sporting an aerodynamic component at the end of the tail rather than tip to butt. This is explained by cheetahs in the wild not having easy access to either carbon fiber or polyethylene, and by a design that the researchers optimized to maximize drag while minimizing mass rather than for biomimicry. “We experimented with a whole array of furry tails to mimic cheetah fur, but found that the half cylinder shape had by far the most drag,” first author Joseph Norby told us in an email. “And the reduction of the drag component to just the end of the tail was a balance of effectiveness and rigidity—we could have made the drag component cover the entire length, but really the section near the tip produces most of the drag, and reducing the length of the drag component helps maintain the shape of the tail.”

Aerodynamic tails are potentially appealing because unlike inertial tails, the amount of torque that they can produce doesn't depend on how much they weigh, but rather with the velocity at which the robot is moving: the faster the robot goes, the more torque an aerodynamic tail can produce. We see this in animals, too, with fluffy tails commonly found on fast movers and jumpers like jerboas and flying squirrels. This offers some suggestion about what kind of robots could benefit most from tails like these, although as Norby points out, the greatest limitation of these tails is the large workspace required for the tail to move around safely.

Tails A variety of animals (and one robot) with aerodynamic drag tails, including a jerboa and giant Indian squirrel. Image: Norby et al

While this paper is focused on quantifying the effects of aerodynamic drag on robotic tails, it seems like there’s a lot of potential for some really creative designs—we were wondering about tails with adjustable floofitude, for example, and we asked Norby about some ways in which this research might be extended. 

I think a foldable or retractable tail would greatly improve practicality by reducing the workspace when the tail is not needed. Essentially all of the animals we studied had some sort of flexibility to their tails, which I believe is a crucial property for improving both practicality and durability. In a similar vein, we've also thought about employing active or passive designs that could quickly modify the drag coefficient, whether by furling and unfurling, or simply rotating an asymmetric tail like our half cylinder. This could perhaps allow new forms of control similar to paddling and feathering a canoe: increasing drag when moving in one direction and reducing drag in the other could allow for more net control authority. This would be completely impossible with an inertial tail, which cannot do work on the environment.

Cheetah tail Gratuitous cheetah picture. Photo: Evan Ackerman/IEEE Spectrum

Even though animals had the idea for lightweight aerodynamic drag tails first, there’s no reason why we need to restrict ourselves to animal-like form factors when leveraging the advantages that tails like these offer, or indeed with the designs of the tails themselves. Without a mass penalty to worry about, why not put tails on any robot that has trouble keeping its balance, like pretty much every bipedal robot, right? Of course there are plenty of reasons not to do this, but still, it’s exciting to see this whole design space of aerodynamic drag tails potentially open up for any robot platform that needs a little bit of help with dynamic motion.

Enabling Dynamic Behaviors With Aerodynamic Drag in Lightweight Tails, by Joseph Norby, Jun Yang Li, Cameron Selby, Amir Patel, and Aaron M. Johnson from CMU and the University of Cape Town is published in IEEE Transactions on Robotics.

The Conversation (0)

How the U.S. Army Is Turning Robots Into Team Players

Engineers battle the limits of deep learning for battlefield bots

11 min read
Robot with threads near a fallen branch

RoMan, the Army Research Laboratory's robotic manipulator, considers the best way to grasp and move a tree branch at the Adelphi Laboratory Center, in Maryland.

Evan Ackerman
LightGreen

This article is part of our special report on AI, “The Great AI Reckoning.

"I should probably not be standing this close," I think to myself, as the robot slowly approaches a large tree branch on the floor in front of me. It's not the size of the branch that makes me nervous—it's that the robot is operating autonomously, and that while I know what it's supposed to do, I'm not entirely sure what it will do. If everything works the way the roboticists at the U.S. Army Research Laboratory (ARL) in Adelphi, Md., expect, the robot will identify the branch, grasp it, and drag it out of the way. These folks know what they're doing, but I've spent enough time around robots that I take a small step backwards anyway.

The robot, named RoMan, for Robotic Manipulator, is about the size of a large lawn mower, with a tracked base that helps it handle most kinds of terrain. At the front, it has a squat torso equipped with cameras and depth sensors, as well as a pair of arms that were harvested from a prototype disaster-response robot originally developed at NASA's Jet Propulsion Laboratory for a DARPA robotics competition. RoMan's job today is roadway clearing, a multistep task that ARL wants the robot to complete as autonomously as possible. Instead of instructing the robot to grasp specific objects in specific ways and move them to specific places, the operators tell RoMan to "go clear a path." It's then up to the robot to make all the decisions necessary to achieve that objective.

Keep Reading ↓ Show less